Answering area acquisition and electronic job processing method, system, equipment and medium
Patent Information
- Application Number
- CN202480001473.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-06-16
- Filing Date
- 2024-03-07
- Publication Date
- 2024-11-12
AI Technical Summary
Existing electronic homework processing technology cannot process and manage student homework efficiently and rationally, resulting in the inability to teach students in accordance with their aptitude on a large scale. The overall teaching quality is affected, and the lack of digital transformation limits the development of the education industry.
Provides a method for obtaining the answer area and processing electronic homework. By determining the question to be processed, obtaining the question attribute information, judging whether it is a handwritten question type, and generating a matching handwritten answer area in the preset blank canvas, automatic recognition and Automatically generate answer areas and support multiple input methods to improve user experience and processing efficiency.
It realizes the electronic operation of handwritten questions, improves the processing efficiency and teaching quality of electronic homework, supports personalized tutoring, helps the education industry realize the digital transformation of teaching, and promotes educational equity and high-quality development.
Smart Images

Figure CN118946910A_ABST
Abstract
Description
Method, system, device and medium for obtaining answer area and processing electronic homework
[0001] This application claims priority to Chinese Patent Application No. 2023102322545, filed on March 10, 2023, and to PCT Patent Application No. PCT / CN2023 / 100845, filed on June 16, 2023. The entire text of the aforementioned Chinese and PCT patent applications is incorporated herein by reference. Technical Field
[0002] The present disclosure relates to the field of data processing technology, and in particular to a method, system, device and medium for obtaining an answer area and processing an electronic homework. Background Art
[0003] Traditional homework is generally completed in paper exercise books, and teachers correct the paper exercise books; the limitations are: 1. It is impossible to handle and manage student homework efficiently and rationally, such as time-consuming typesetting and printing of test papers, and test papers are easy to lose, etc.; 2. It is impossible to know the individual learning situation of students, so the students themselves, parents, and teachers cannot quantify the student's learning situation; in the case of class lessons, it is impossible to assign homework to students in a class that targets the specific weaknesses of each student, nor is it possible to further challenge the top students in the subject with more difficult homework, and large-scale teaching in accordance with their aptitude cannot be achieved; 3. School subject groups, education authorities, etc. are unable to monitor academic quality; education authorities and responsible parties cannot guarantee the quality and quantity of assigned homework, and teachers assign a large amount of homework for the sake of exam results. Too much homework causes students to have a heavy academic burden, and there is little homework assigned, and it is impossible to supervise the quality of the homework; the education authorities and responsible parties are unable to supervise teachers' grading of homework, resulting in situations such as teachers asking students to grade each other's homework; the education authorities and responsible parties are unable to monitor whether the content of homework exceeds the syllabus, and are unable to monitor whether the teaching content violates ideology; 4. Due to the lack of process data, the core link of teaching in the education industry has not been able to be digitized so far, which further leads to the inability to release the value potential of commercial data; 5. Due to the inability to digitize the teaching process, the education industry has not been able to further develop the application of artificial intelligence like other industries; 6. Promote the digital transformation of education and promote fair, balanced and high-quality development of regional education through technological means.
[0004] Currently, there are some attempts at electronic grading beyond the traditional paper-based approach. These include: One approach involves inputting multiple-choice questions and automatically grading them. However, the technical limitations of this approach lie in the fact that multiple-choice questions only account for a small portion of the assignments in subjects like Chinese, mathematics, English, physics, and chemistry. The vast majority of assignments are complex, such as essay questions and fill-in-the-blank questions. Therefore, automated grading of multiple-choice questions alone is not widely applicable. Alternatively, students can upload photos of their responses to each question, which can be inconvenient, time-consuming, and provide a poor user experience. Furthermore, it also lacks the ability to analyze student performance.
[0005] Summary of the Invention
[0006] The technical problem to be solved by the present disclosure is to overcome the various drawbacks of electronic homework in the existing technology that are generally unable to efficiently and rationally process and manage student homework, resulting in the inability to teach students in accordance with their aptitude on a large scale and affecting the overall teaching quality. A method, system, device and medium for obtaining answer areas and processing electronic homework are provided.
[0007] The present disclosure solves the above technical problems through the following technical solutions:
[0008] The present disclosure provides a method for obtaining an answer area, which is applied in electronic homework and includes:
[0009] Identify pending issues;
[0010] Obtaining the topic attribute information of the topic to be processed;
[0011] Determining whether the pending question is a handwritten question type based on the question attribute information, and if so, generating a handwritten answer area matching the pending question in a first preset blank canvas;
[0012] The first preset blank canvas is a carrier for handwriting input content.
[0013] The present disclosure further provides a method for processing an electronic homework, which is implemented according to the above-mentioned method for obtaining the answer area, and includes:
[0014] Based on the preset learning objectives, select several questions from the preset question bank;
[0015] Generate a handwritten answer area corresponding to each question of the handwritten question type using the acquisition method;
[0016] Generate set answer positions corresponding to other remaining questions of non-handwritten question types;
[0017] The electronic homework is obtained based on the question and the corresponding handwritten answer area and the set answer position.
[0018] The present disclosure also provides a system for acquiring a handwritten answer area, which is applied in electronic homework and includes:
[0019] A topic determination module is used to determine the topic to be processed;
[0020] An attribute information acquisition module, configured to acquire the topic attribute information of the topic to be processed;
[0021] A first judgment module is used to judge whether the to-be-processed question belongs to the handwriting question type based on the question attribute information, and if so, call the area generation module;
[0022] The area generation module is used to generate a handwritten answer area matching the question to be processed in the first preset blank canvas;
[0023] The first preset blank canvas is a carrier for handwriting input content.
[0024] The present disclosure further provides a system for processing electronic homework, which is implemented based on the above-mentioned system for obtaining the handwritten answer area, and includes:
[0025] The topic selection module is used to select a number of topics from a preset topic library based on preset learning objectives;
[0026] An answer area determination module is used to generate a handwritten answer area corresponding to each question of the handwritten question type using the acquisition system;
[0027] The answer position generation module is used to generate set answer positions corresponding to the remaining questions of the non-handwritten question type;
[0028] The electronic homework acquisition module is used to acquire the electronic homework based on the question and the corresponding handwritten answer area and the set answer position.
[0029] The present disclosure further provides an electronic operation operation and management platform, which includes the electronic operation processing system described above and a plurality of preset user terminals for communication interaction;
[0030] The plurality of preset user terminals include a student homework application terminal in a student device and a teacher homework application terminal in a teacher device.
[0031] The present disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, the above-mentioned method for obtaining the answer area is implemented; or, the above-mentioned method for processing electronic homework is implemented.
[0032] The present disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for obtaining the answer area; or, implements the above-mentioned method for processing electronic homework.
[0033] On the basis of conforming to the common sense in this field, the above-mentioned preferred conditions can be arbitrarily combined to obtain the preferred embodiments of the present disclosure.
[0034] The positive progress of this disclosure is:
[0035] In the present disclosure, the functions of automatically identifying any handwritten question in an electronic homework and automatically generating an answer area enable a series of electronic operations on handwritten questions to be realized, including but not limited to: (1) enabling a single electronic homework containing an answer area for a handwritten question to be realized; (2) enabling a collection of several electronic homeworks containing handwritten questions, namely an electronic homework book (such as several mathematical electronic homeworks constituting a digital electronic homework book); (3) the realization of the electronic homework book enables the subversive realization and optimization of the effectiveness of various educational scenarios, including but not limited to: teachers improve their own work efficiency through electronic homework, such as through electronic homework, the system can help to count the personalized learning situation of students; thereby facilitating teachers to provide targeted tutoring to different students; students can better achieve their learning goals through electronic homework, such as automatically collecting wrong questions and pushing exercises, etc.; giving specific units Organize supervision and management of schools and teachers to provide information support for the teaching process, such as checking the amount of assigned homework; help the education industry realize the digital transformation of teaching, a core link, and further release the value potential of commercial data; benefit from the digital transformation of the teaching process, the education industry can continue to explore the application of artificial intelligence like other industries; promote the digital transformation of education, and promote fair, balanced and high-quality development of regional education through technical means; that is, in this embodiment, by automatically identifying and processing any pending questions in the electronic homework, and generating a handwritten answer area for the identified handwritten questions, no human intervention is required in the process, and it is all automated, which improves the overall processing efficiency and electronic level of electronic homework, and also improves the user experience; in addition, it provides reliable guarantees for large-scale teaching according to students' aptitude and comprehensively and high-quality improvement of teaching quality. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] FIG1 is a flow chart of a method for obtaining an answer area in Example 1 of the present disclosure.
[0037] FIG2 is a schematic diagram of the answer area of the handwriting question in Example 2 of the present disclosure.
[0038] FIG3 is a first schematic diagram of the operation layer of the handwriting answering area in the second embodiment of the present disclosure.
[0039] FIG4 is a second schematic diagram of the operation layer of the handwriting answering area in the second embodiment of the present disclosure.
[0040] FIG5 is a third schematic diagram of the operation layer of the handwriting answering area in the second embodiment of the present disclosure.
[0041] FIG6 is a schematic diagram of a first display state of a handwriting question according to Embodiment 2 of the present disclosure.
[0042] FIG7 is a schematic diagram of a second display state of a handwriting question according to Embodiment 2 of the present disclosure.
[0043] FIG8 is a schematic diagram of a third display state of the handwriting question according to the second embodiment of the present disclosure.
[0044] [Corrected 19.04.2024 according to Rule 91] Figure 8a is a first schematic diagram of the answer area of Example 2 of the present disclosure.
[0045] [Corrected 19.04.2024 according to Rule 91] Figure 8b is a second schematic diagram of the answer area of Example 2 of the present disclosure.
[0046] [Corrected 19.04.2024 according to Rule 91] Figure 8c is a third schematic diagram of the answer area of Example 2 of the present disclosure.
[0047] [Corrected 19.04.2024 according to Rule 91] Figure 8d is a fourth schematic diagram of the answer area of Example 2 of the present disclosure.
[0048] [Corrected 19.04.2024 according to Rule 91] Figure 8e is a fifth schematic diagram of the answer area of Example 2 of the present disclosure.
[0049] [Corrected 19.04.2024 according to Rule 91] Figure 8f is a sixth schematic diagram of the answer area of Example 2 of the present disclosure.
[0050] FIG9 is a flow chart of a method for obtaining an answer area in Embodiment 2 of the present disclosure.
[0051] FIG10 is a schematic diagram of the initial size of the handwriting answer area in Example 2 of the present disclosure.
[0052] FIG11 is a first extended display diagram of the handwriting answer area in Example 2 of the present disclosure.
[0053] FIG12 is a second extended display diagram of the handwriting answer area in Example 2 of the present disclosure.
[0054] FIG13 is a first schematic diagram of a basic function plug-in of the handwriting answering area in Example 2 of the present disclosure.
[0055] FIG14 is a second schematic diagram of the basic function plug-in of the handwriting answering area in Example 2 of the present disclosure.
[0056] FIG15 is a schematic diagram of the first scenario of calling a draft paper in the handwriting answering area according to the second embodiment of the present disclosure.
[0057] FIG16 is a schematic diagram of a second scenario of calling a draft paper in the handwriting answering area according to the second embodiment of the present disclosure.
[0058] FIG17 is a flowchart of a method for processing electronic tasks according to a third embodiment of the present disclosure.
[0059] FIG18 is a first flow chart of the method for processing electronic tasks according to the fourth embodiment of the present disclosure.
[0060] FIG19 is a second flow chart of the method for processing electronic tasks according to the fourth embodiment of the present disclosure.
[0061] FIG20 is a third flow chart of the method for processing electronic tasks according to the fourth embodiment of the present disclosure.
[0062] FIG21 is a fourth flow chart of the method for processing electronic tasks according to the fourth embodiment of the present disclosure.
[0063] FIG22 is a fifth flow chart of the method for processing electronic tasks according to the fourth embodiment of the present disclosure.
[0064] FIG23 is a sixth flow chart of the method for processing electronic tasks according to the fourth embodiment of the present disclosure.
[0065] Figure 24 is a schematic diagram of a storage data mapping relationship table according to embodiment 4 of the present disclosure. Figure 25 is a schematic diagram of a question bank folder according to embodiment 4 of the present disclosure.
[0066] FIG26 is a first schematic diagram of homework folders of different classes according to the fourth embodiment of the present disclosure.
[0067] FIG27 is a second schematic diagram of homework folders for different classes according to the fourth embodiment of the present disclosure.
[0068] FIG28 is a schematic diagram of the setting interface of the operation mode of Example 4 of the present disclosure.
[0069] FIG29 is a schematic diagram of the setting interface of the examination mode of Example 4 of the present disclosure.
[0070] FIG30 is a schematic diagram of setting a time rule in Example 4 of the present disclosure.
[0071] FIG31 is a schematic diagram showing the answers of different students to the same question in Example 4 of the present disclosure.
[0072] Figure 32 is a schematic diagram of the interface of the wrong question book of embodiment 4 of the present disclosure.
[0073] FIG33 is a module diagram of the system of the answering area of Example 5 of the present disclosure.
[0074] FIG34 is a module diagram of the system of the answering area of Example 6 of the present disclosure.
[0075] FIG35 is a module diagram of the electronic operation processing system of Example 7 of the present disclosure.
[0076] FIG36 is a module diagram of the electronic operation processing system of Example 8 of the present disclosure.
[0077] FIG37 is a schematic structural diagram of an electronic device according to embodiment 10 of the present disclosure. DETAILED DESCRIPTION
[0078] The present disclosure is further illustrated below by way of examples, but the present disclosure is not limited to the scope of the examples.
[0079] In the current educational landscape, there are widespread drawbacks in the inability to efficiently and rationally handle and manage student homework. Specifically, the following three solutions are commonly used to handle and manage homework:
[0080] Solution A: Manually print out the entire assignment in paper answer boxes in advance. Students write their answers in each paper answer box and then scan the completed answer box to obtain a scanned image. The scanned image is then uploaded to a designated system. The teacher views the scanned image in the designated system and corrects the answers.
[0081] The disadvantages of this solution are that the acquisition efficiency and degree of automation of the entire electronic operation are low, the operation process is cumbersome, and the user experience is poor.
[0082] Solution B: Use special dot-matrix paper to demarcate the answer area for each question, and use a matching dot-matrix electronic pen to answer the questions. The dot-matrix paper and dot-matrix electronic pen are connected to a Bluetooth module, and a computer running software that receives Bluetooth signals is required to complete the collection of answers in the electronic homework.
[0083] The disadvantages of this solution are: the hardware investment cost and software investment cost are both high, and although the student answers are accessible, it is impossible to match different answers with specific questions, that is, the questions and students' answers are separated from each other, and the data management dimension is single, making it impossible to complete personalized learning statistics.
[0084] Plan C: After students answer the questions on paper, they take a photo of their handwritten answers for each question and upload the photos to a designated system. The teacher then checks the photos in the designated system and corrects the answers.
[0085] The disadvantages of this solution are: students generally resist the additional operation of taking photos and uploading them, which means that the student experience is poor; the font sizes of the uploaded pictures are different, and it is time-consuming and laborious for teachers to correct them, which results in a poor user experience for teachers, and then leads to low efficiency and automation in obtaining the entire electronic homework, which cannot meet the high-demand teaching scenarios.
[0086] On this basis, the present disclosure innovatively proposes a method for obtaining an answer area and a method for processing electronic homework to solve various problems commonly found in the above-mentioned homework processing solutions. Specifically:
[0087] Example 1
[0088] As shown in FIG1 , this embodiment provides a method for obtaining an answer area, which is applied in electronic homework and specifically includes the following steps:
[0089] S101, determining the problem to be processed;
[0090] Among them, the questions to be processed are generally selected from a preset question library.
[0091] S102, obtaining the topic attribute information of the topic to be processed;
[0092] Specifically, the question attribute information includes but is not limited to question content and question type information.
[0093] S103, judging whether the question to be processed belongs to the handwriting question type based on the question attribute information, if it does, executing step S104;
[0094] S104: Generate a handwritten answer area matching the question to be processed in the first preset blank canvas;
[0095] The first preset blank canvas is a carrier for handwriting input content.
[0096] Specifically, the handwriting question type is a type that requires students to answer questions by hand; the preset blank canvas is a carrier for users to input handwriting content.
[0097] For example, for question A, if it is determined to be a handwritten question based on the question attributes, a handwritten answer area will be generated promptly and automatically in the blank canvas below question A. The answer area is an answer box for students to answer question A, as shown in Figure 2 (the content of the question stem in the figure is only for illustrative purposes and is applicable to any handwritten question), and S corresponds to the handwritten answer area; if question A is not a handwritten question, the above logical steps do not need to be executed.
[0098] The above-mentioned method for obtaining the answer area can be applied on the teacher side, the student side, or both the teacher side and the student side to meet the operation requirements of the teacher side and the student side for the questions respectively.
[0099] Among them, the following methods can be used to determine the question type:
[0100] Method (1): manually label each question in the question bank in advance, so that during the automatic question type determination process, the question label can be directly obtained to identify the question type;
[0101] Method (2): Convert the document information in Word or PDF into code format and scan it in sequence, judging based on the context of the original paper, such as "I. Multiple Choice Questions", "II. True or False Questions", "III. Answer Questions", etc.
[0102] Among them, PDFs, pictures, etc. are scanned to obtain the original electronic homework data, the homework data is analyzed, and the question numbers of different levels in the homework data are extracted. There are one or more question numbers under the same level;
[0103] The specific implementation method of extracting the question number in the big question: determine whether there are pre-set question numbers of the type in the document, such as ("one," "two," "three," ..., "1.", "2.", "3." ..., "(1).", "(2).", "(3)."), if so, extract and mark them; through the hierarchical relationship between the pre-set question numbers, such as "one," "two," "three," corresponding to the first level, 1.", "2.", "3." corresponding to the second level, "(1).", "(2)." (3) corresponding to the third level, and the actual question number, automatically obtain each sub-topic under the big question. For any big question, extract the number of question numbers in the big question and determine how many sub-topics there are; for each sub-topic, determine whether it is a handwritten question;
[0104] Specifically, determine whether it is a multiple-choice question: Specifically determine whether there is a certain distance between each two letters, and there is content in the middle, and the answer format is a single letter. If all the above conditions are met, it is generally judged to be a multiple-choice question; if it is a multiple-choice question, first exclude:
[0105] Determine whether it is a handwritten question: Determine whether there are underscores or brackets (spaces between the left and right brackets) in the question. If so, it is determined to be a handwritten question;
[0106] Determine whether the question consists of a stem and a large blank space (count the number of line breaks in the blank area. If the number of line breaks is greater than a set threshold, it indicates that there is a large blank space). If so, it is probably a handwritten question.
[0107] Judge based on the answer content: If the answer content is T / F, then the question is judged to be a true / false question, and handwriting questions are excluded; if the answer content is options, then the question is judged to be a multiple-choice question, and handwriting questions are excluded; if the answer content is a short or long rich text format (such as composed of text, pictures, formulas, etc.), then it is judged to be a fill-in-the-blank question or a complex answer question.
[0108] Method (3): The specific process of determining whether it is a handwritten question type through the algorithm model is as follows:
[0109] Preprocessing: Convert the question content into rich text code format. The original code may contain text, formulas, and images. Formulas, etc. are stored in a coding format (such as LaTeX format). Images are first converted to black and white and saved in PNG format (a type of image format). This losslessly preserves the original pixel information and records the relative position of the image in the question. For the text portion, delete any information that is not useful for handwritten question judgment, such as formatting information (font, size, color, etc.).
[0110] Due to the wide variety of question formats, such as the text features like "ABCDE" mentioned in the above method, it is inevitable that some multiple-choice questions will use 12345, abcde, etc. as option labels. Using conditional judgments simply cannot exhaust all possibilities. However, using AI, based on data pairs of existing questions and manually labeled data, can train models to learn more abstract and implicit features of questions with the same label (such as the presence of options in multiple-choice questions and the frequent "explanation" requirements or questions in short-answer questions). This allows for better identification of question styles and question structures that manual rules cannot account for.
[0111] Text feature extraction: Using pre-trained semantic models or sentence embedding models (such as Sentence BERT) and other models, with the existing pre-trained Masked Language Model (MLM, taking the BERT model as an example), since BERT uses general text for pre-training, it is necessary to fine-tune it on the answer data so that the model can obtain feature extraction that is more relevant to the content of the homework questions by learning the joint distribution of various mathematical symbols, proper nouns and other words. Among them, the fine-tuning operation includes but is not limited to: using the questions with random words masked as the model input, and then training the model to predict the originally masked words; the masked words are known to be part of the question text or a mathematical formula, so the model is trained to predict the type of masked words at the same time; the known question metadata (subject, grade, etc.) is used as a multi-task learning goal, and the model is trained to predict these metadata of the questions at the same time; using this model, the question content is subsequently mapped to a high-dimensional space to obtain a high-dimensional vector as the text semantic feature of the question;
[0112] Specifically, the BERT model first splits the text into characters, words, and symbols to obtain several minimum semantic units. Each semantic unit is converted into a numerical index according to a pre-set mapping table. At the same time, a special unit [CLS] is added to the front of the question to prepare for subsequent question classification. The semantic units are converted into high-dimensional vectors using a fine-tuned BERT model. The vector of the first token [CLS] of the question is used as the feature vector of the entire question. That is, a neural network (such as transformer) or a word vector model (such as word2vec) is used to convert the semantic units into high-dimensional vectors. The vectors of all the semantic units in the entire question are then integrated (e.g., averaged) to obtain a high-dimensional vector of the entire question, which is used as the text semantic feature of the question.
[0113] Image feature extraction: Use a convolutional neural network (CNN) to map the question content into a high-dimensional space and obtain a high-dimensional vector as the image feature of the question; specifically, the numerical value corresponding to each pixel of the image (if it is a black and white image, it is a single integer in the range of 0-255, representing the intensity of white; if it is a color image, it is a 3-dimensional vector, and each numerical value is also in the range of 0-255, representing the intensity of the three primary colors of red, green, and blue respectively); based on this, a matrix (for black and white images, an m by n matrix, where m and n are the pixel length and width of the image, and the value of each matrix represents the intensity of white) or a tensor (for color images, an m by n matrix, where m and n are the pixel length and width of the image, and the value of each matrix is a 3-dimensional vector, representing the intensity of red, green, and blue respectively) is generated. This matrix / tensor can be used as the input of the convolutional neural network to extract a high-dimensional vector as the image feature of the question;
[0114] Alternatively, based on the pre-trained Vision Transformer model (an image classification model), fine-tuning is performed on the question image. The fine-tuning operation includes: dividing the question image, masking random words as model input, and then training the model to predict the originally masked image segments; since the question image may have large blank areas, only the segments with non-blank parts reaching a certain threshold are selected for masking and training; known question metadata (question type, such as trigonometric function questions, limits, etc.) are used as multi-task learning objectives, and the model is trained to predict these metadata of the questions at the same time; or joint fine-tuning and comparative training are adopted: features are extracted from a pair of questions belonging to different categories respectively, and the feature extractor is fine-tuned to make the similarity (Euclidean distance) of the extracted features as large as possible, thereby enhancing the feature extractor's ability to extract question category information and discrimination.
[0115] Training and prediction: The image and text feature extractors will extract features (n-dimensional vectors) respectively. If the image does not exist, a vector of all zeros will be used instead. The two vectors are concatenated to obtain a vector of length 2n, which is then input into a single-layer neural network to output a vector of length m. Each number in the vector can obtain the probability of the type of question it represents through the Softmax function (normalized exponential function), where m is the number of question types. During training, the cross-entropy loss is calculated to maximize the probability of the model predicting the correct type of question; at the same time, an additional loss function is added to minimize the probability of the wrong type of question, thereby reducing the loss until convergence; when using the model to predict the question type, the question feature vector is input, and the question type with the highest probability is selected as the model's prediction result;
[0116] Among them, the prediction results will be manually checked, and the model training data will be regularly updated according to the inspection results to continuously optimize the model; or when the system is put into use, the question type calculated by the model will be provided as the default value to the users who use this question. If for a small number of questions, the user manually modifies the question type before starting to use this question, it means that the model judgment is incorrect. Based on this error, the model parameters are retrained, and the model training data is updated based on this error to reduce errors, thereby improving the accuracy of the model prediction results and meeting the usage requirements of actual work scenarios.
[0117] It should be noted that: when the judgment results given by different judgment methods in the above method (2) are inconsistent, the model in method (3) is called for auxiliary confirmation; if necessary, the analysis results obtained under each method can be sent to the corresponding personnel for confirmation to effectively ensure the accuracy of the handwriting question determination.
[0118] In addition, the data structure of the question is relatively detailed, and the text data, options, answer scores and other data corresponding to the question stem are all stored independently; according to the answer format and content, a specific algorithm model can be used for training to directly predict whether the question is a multiple-choice question or a judgment question; the acquisition of the model can be trained according to actual needs, which will not be repeated here. In this embodiment, considering that there are significant differences between the specific question features of handwritten questions and non-handwritten questions, the corresponding question features are identified through the question attribute information to make timely and effective distinctions between the two types of questions; it does not emphasize which specific recognition technology is used for processing, as long as it can accurately analyze whether it is a handwritten question. For handwritten questions, an answer area corresponding to the handwritten question can be automatically generated. The answer area can be used for students to write their answers by hand, and can also be used for teachers to correct students' answers.
[0119] In this embodiment, the function of automatically identifying any handwritten question in the electronic homework and automatically generating the answer area enables a series of electronic operations on the handwritten question, including but not limited to:
[0120] (1) Enables the creation of a single electronic assignment containing a handwritten answer area;
[0121] (2) It enables the collection of several electronic homework sets containing handwritten exercises, i.e., electronic homework books (e.g., several sets of mathematics electronic homework sets constitute a digitized electronic homework book);
[0122] (3) The implementation of electronic homework books enables the subversive realization and optimization of the effectiveness of various educational scenarios, including but not limited to: (a) teachers improve their work efficiency through electronic homework, such as through electronic homework, the system can help count the individual learning situation of students; (b) it is convenient for teachers to provide targeted tutoring to different students; students can better achieve their learning goals through electronic homework, such as automatically collecting wrong questions and pushing personalized exercises; (c) it provides information support for specific units to organize supervision and management of schools and teachers' teaching processes, such as checking the amount of homework assigned; (d) it helps the education industry realize the digital transformation of teaching, a core link, and further release the value potential of commercial data; (e) thanks to the digital transformation of the teaching process, the education industry can continue to explore the application of artificial intelligence like other industries; (f) it promotes the digital transformation of education and promotes fair, balanced and high-quality development of regional education through technological means.
[0123] Therefore, the automatic generation of the answer area for any handwritten question in the processing questions is a prerequisite for optimizing different educational scenarios and making them possible.
[0124] That is to say, in this embodiment, any pending questions in the electronic homework are automatically identified and processed, and an answer area for handwriting is generated for the identified handwritten questions. No human intervention is required in the process, and it is all automated, which improves the overall processing efficiency and electronic level of the electronic homework, and also improves the user experience.
[0125] Example 2
[0126] The method for obtaining the answer area in this embodiment is a further improvement on the solution in Example 1. Specifically:
[0127] In an implementable solution, the first preset blank canvas is an initial layer generated based on a canvas in a picture format; wherein the picture format includes but is not limited to PNG and JPG (PNG and JPG are both picture formats).
[0128] In this solution, once a question is identified as a handwritten question, an initial layer for handwritten answers is automatically generated. This initial layer can be obtained in a timely manner without any other unnecessary processing. After the student completes the answer in the answer area of the initial layer and submits the electronic homework, the teacher can automatically obtain a new operation layer when opening the electronic homework. In the subsequent process, each time the student or teacher enters the electronic homework, a new operation layer will be generated based on the previous one.
[0129] It should be noted that the initial layer allows students to have an electronic answer sheet for their initial assignments, that is, from the beginning of the entire assignment process, the students' assignments are electronic and in one place, thus realizing unified electronic management of assignments.
[0130] The solutions in the prior art are: either the questions are printed on paper, and the students write their answers on paper, which are then uploaded to the computer later. In this case, the students' written homework and the archived homework are separated in two places; or the students write their answers on special dot paper, and the writing traces are uploaded to the computer. In this case, the students' written homework and the archived homework are separated in two places; or the students write their answers on ordinary paper homework books, which are uploaded to the computer by taking a photo. In this case, the students' written homework and the archived homework are separated in two places. In other words, the above-mentioned existing solutions generally have the defects of separating the students' written homework and the archived homework, being unfriendly to users, and having low processing efficiency. The solution disclosed in this disclosure can effectively ensure the convenience and flexibility of answering questions in the answering area and facilitate electronic management.
[0131] In an implementable solution, the handwritten answer area is used to display the corresponding stacked content in all completed and newly generated operation layers that are stacked in sequence;
[0132] Among them, all the operation layers stacked in sequence are generated in sequence when different users enter the handwriting answering area at different preset usage stages.
[0133] The trajectory process in each operation layer is stored separately, and each layer's trajectory can be analyzed separately. The student's answer and correction data in different layers can be analyzed individually or as a whole, achieving the goal of controlling the monitoring of multiple aspects of the student's handwriting. Similarly, the teacher's correction data in different layers can be analyzed individually or as a whole, achieving the goal of controlling the monitoring of multiple aspects of the teacher's marking process.
[0134] Specifically, if you need to know how seriously the teacher team is grading homework: mainly by analyzing the time spent by teachers on grading, and extracting the trajectory data of each teacher's grading layer separately, you can count what percentage of questions the teacher has simple marks of ticking or crossing, and what percentage of questions the teacher spent more than a preset threshold (such as 5 seconds) to write the specific reasons for the error, and give the teaching management personnel a comprehensive data-based indicator of grading quality; you can also break it down into subject groups and individual teachers, and look at the grading seriousness indicators of each subject group and individual teacher; if you need to analyze whether students take notes when listening to the lectures and test papers: you can extract the students' trajectory data on the note layer separately, and count what percentage of wrong questions the students have added effective notes.
[0135] For people in the homework system (with operating permissions), after opening the electronic homework, they can trigger a new layer and write content in the layer, such as by clicking on a writing pen, eraser, etc.; for each operator, a corresponding layer is automatically generated in the answer area of the electronic homework they have recently obtained. The current operator only has the permission to operate the current layer, and all data corresponding to the operation will be recorded in the layer to ensure the independence of the data of each layer; and the current operator does not have any port or way to interfere with other layers. In addition, each completed layer is set to be unable to be modified, thus ensuring the reliability of each layer of data. In addition, for exported homework (such as the answer content of a student to a certain question shared with parents), no new layer will be triggered and cannot be traced back. Specifically, as shown in Figure 3, the operation layers generated at different preset usage stages include an initial layer for the student side, a correction writing layer for the teacher side, a correction layer for the student side, or a correction correction layer for the teacher side;
[0136] Among them, according to different scene requirements, the number of layers contained in the operation layer can be different. For example, some scenes only involve the initial layer and the correction writing layer, and some scenes only involve the initial layer, the correction writing layer and the correction layer, etc.; no matter how many layers the operation layer contains, these operation layers are stacked in the order of initial layer, correction writing layer, correction layer, and correction layer.
[0137] In addition, after N correction operations, a correction layer for the student side and a correction layer for the teacher side are generated again in the handwritten answer area, where N is a positive integer.
[0138] At this time, as shown in FIG4 , these operation layers are stacked in sequence: initial layer, correction layer, correction layer 1, correction layer 1, correction layer 2, correction layer 2, correction layer 3, correction layer 3, ...
[0139] Specifically, as shown in Figure 5, from left to right are the initial layers used by the student side. After the student side writes content on the initial layer and submits the electronic homework, when the teacher side opens the electronic homework and enters the question, a correction writing layer for the teacher side to use will be automatically formed on the current answer area. After the teacher completes the correction, the merged content of the two operation layers is formed; the implementation process of the merged content obtained by stacking between other layers is similar, so it will not be repeated here.
[0140] Compared with the solutions in the existing technology, the scanned or photographed pictures uploaded by students are used as the student answer layer, and there are no other layers besides the initial layer, such as the correction layer and all subsequent layers; therefore, there are only scores, but many other dimensional information is missing, and the supported functions are relatively simple, and cannot meet the usage needs of higher-demand actual educational scenarios; and, corrections are made in electronic archives, and the electronic archives and the original answer sheets are separated, and the correction content cannot be displayed on the original student workbooks.
[0141] In this solution, there is an operation layer when students answer questions, an operation layer when teachers correct questions, an operation layer when students make corrections after seeing the teacher's corrections, and an operation layer when the teacher corrects questions again. These operation layers are independent of each other to ensure that the operations completed by each do not affect each other, so that the display content of the answer area of each question is independent and complete at different processing stages, which effectively solves the problems existing in the above-mentioned existing solutions.
[0142] In one feasible solution, the acquisition method further includes:
[0143] Get the re-answer instruction corresponding to the pending question;
[0144] Based on the re-answer instruction, a new handwritten answer area matching the question to be processed is generated in the second preset blank canvas;
[0145] The properties and functions of the new handwriting answer area are completely different from those of the handwriting answer area previously generated in step S104 . The only difference is that they are generated on different blank canvases.
[0146] Control all operation layers in the handwriting answer area obtained based on the first preset blank canvas to be in a display state or a hidden state.
[0147] Specifically, by creating a new array in the system's database and distinguishing it from the original array (storing the first preset blank canvas), each item in the newly created array is the stackable canvas mentioned above. The advantage of this solution over not adding a new canvas (i.e. correcting on the original canvas) is that: for example, with the correction track, the track on the initial canvas can still be extracted separately, which is convenient for students, teachers, and parents to understand the first answer situation and why they made mistakes; the track on the correction canvas can also be extracted separately, and only the track on the correction canvas can be viewed without being disturbed by the old content; the student's correction habits can also be analyzed, and among all the wrong questions, how many new answer areas have been generated and the questions have been corrected; the difference between the student's first answer and the second answer can also be analyzed, and how much time has been shortened for each step, etc., so as to achieve the purpose of obtaining more dimensional learning situation analysis.
[0148] In this solution, the user has already completed a short-answer question within a certain period of time. That is, the answer area for the short-answer question and multiple operation layers within the answer area have been generated. When the user needs to answer (or correct) the short-answer question again, such as when reviewing or recalling the question, they need to perform specific operations on the current short-answer question, such as clicking the first specific button set under the current short-answer question to trigger the re-generation of a new answer area corresponding to the short-answer question.
[0149] The user can click the second specific button of the current short-answer question and choose whether to display or hide the operation layer in the answer area obtained based on the first preset blank canvas; when choosing to hide, the user can answer the question without interference; when choosing to show, the user can refer to the previous answer and correction status when answering the question or after answering the question to understand his or her own level of mastery of the current short-answer question.
[0150] Among them, when choosing to display the operation layer in the answer area obtained based on the first preset blank canvas, the new answer area can be generated before the answer area generated in the past, or after the answer area generated in the past, or the new answer area and the answer area generated in the past can be arranged side by side. Of course, there can be other ways to display and arrange them. The specific method to be adopted can be determined or adjusted according to actual needs.
[0151] In addition, the answer area of the short-answer question can be regenerated for the third, fourth, etc. time, so that the user can practice answering the same short-answer question multiple times. The implementation process is similar to the above case and will not be repeated here.
[0152] In addition, in the existing scheme, when students answer the same question again, it is impossible to establish the relationship that the multiple answers correspond to the same question, and they are separated from each other; for example, when students make corrections, if they do not write on the original paper, they need to write on another piece of paper. At this time, the corrected paper and the original paper are separated, and are separated from the same question; if students do not write the corrections on the original page of the homework book, they will write on another page of the homework book. Then the corrected answer page and the original answer page are not in the same place. When reviewing in the future, they will need to search back and forth, which is bound to cause operational problems; while in this scheme, the answer canvas or answer area during correction is always displayed together with the previous historical answer area, thereby ensuring that the multiple generated answer areas for the same question are always closely related, while satisfying the user's correction needs, it is convenient for unified management of the content in different answer areas under the same question.
[0153] In one feasible solution, the handwriting answering area is an area for responding to and displaying handwriting touch content; and / or an area for responding to and displaying content input based on an external input device;
[0154] The handwriting answer area is displayed on the display interface of the smart terminal that supports handwriting touch operation.
[0155] Compared with existing solutions, either they do not support electronic handwriting input, or although they are an electronic input method, they are not electronic devices, and there are problems such as inconvenient operation and low implementation flexibility; in this solution, the answer area can support content input by handwriting touch, and can also support content input through external devices (including but not limited to mouse and keyboard), that is, the answer area is compatible with multiple input methods, which ensures the flexibility of users' operation of the answer area and improves the user's answering experience.
[0156] In an implementable solution, the question answering area is an area for displaying handwritten content of handwritten touch content.
[0157] In this solution, the answer area can support real-time display of electronic tracks at the tip of the electronic pen, that is, the input content is output and displayed in real time, ensuring the handwriting experience and avoiding problems such as delayed display that cause a poor user experience.
[0158] In addition, smart terminals include but are not limited to tablets, tablet-keyboard all-in-one computers, computers with handwriting tablets, and touch-screen computers.
[0159] Existing solutions require either handwritten answers on paper, which is then scanned and stored on a computer; traditional pen and paper are used for writing, which is then photographed and uploaded to a computer or mobile phone; or digital pens are used to write on dot paper, where the dot paper remains a physical sheet of paper, with the pen's writing traces transmitted to the computer via Bluetooth. In other words, for handwritten questions, these existing solutions require the original answers to be written on paper first and then transmitted.
[0160] In this solution, an electronic stylus is used to handwrite on the electronic screen of a tablet PAD (tablet computer), which is convenient for users to write; the homework is electronic from the beginning, and there is no need for additional steps such as taking pictures, scanning, and splitting, so that the initial answers and subsequent trajectories can be stored together, displayed in one place, and subsequently processed; the data entered by the user's handwriting on the tablet can be synchronized with the computer and mobile phone, and the handwritten data synchronized from the tablet can be viewed or edited on different devices (tablet, computer or mobile phone) according to the usage requirements of different scenarios, and the handwriting input of different smart terminals does not affect the operation of other smart terminals.
[0161] That is to say, in this solution, the process of determining the answer area of any handwritten question in the electronic homework can be completed in the smart terminal. Preferably, it is implemented in the tablet PAD, which further meets the user's usage habits and improves the user's usage experience.
[0162] In one feasible solution, the electronic homework includes a plurality of stems and a plurality of to-be-processed questions under each stem;
[0163] Step S104 includes:
[0164] For each pending question of the handwritten question type in the electronic homework, a handwritten answer area for each pending question is generated on the first preset blank canvas.
[0165] Specifically, the specific implementation method of generating answer areas for different questions in any large question is: determine whether there are pre-set question numbers of the type in the homework data of the electronic homework, such as ("one," "two," "three," ..., "1.", "2.", "3." ..., "(1).," "(2).," "(3)."), and automatically obtain each pending question under the large question; use the above-mentioned schemes (1), (2), and (3) for determining the question type to determine whether each question is a handwritten question, and if so, generate the corresponding answer area.
[0166] Based on the above method, the answer area corresponding to each handwritten question type is automatically generated on the first preset blank canvas; the written content submitted by the students is stored in different data units, and each data unit is automatically marked with the answer content corresponding to which student submitted it and which specific question. In this solution, an electronic homework generally contains multiple pending questions, and all pending questions are identified and processed in order of question numbers, or multiple pending questions are identified and processed at the same time. As long as one or more pending questions are handwritten questions, the corresponding answer areas are automatically generated for each question. That is, this solution can realize batch identification and processing of multiple pending questions, and can also batch generate matching answer areas for multiple identified handwritten questions, thereby improving the efficiency of electronic homework generation and the level of electronicization.
[0167] In addition, it should be noted that: in the existing technical solutions, either an answer sheet with a split frame is manually designed, and after scanning the entire answer sheet, the answer content corresponding to each question is obtained according to the frame lines; however, the process of assigning homework and doing homework in school is repeated frequently every day, and teachers are not happy to go through the trouble of manually splitting the answer frame for students and then printing and scanning the homework every day. The existing solution can only be used in extremely low-frequency usage scenarios such as the college entrance examination. It is acceptable to split the answer area of a test paper for the college entrance examination only once a year, but it does not have the user experience conditions for high-frequency scenarios; in this solution, because it has the function of automatic system splitting of homework, it is no longer based on manual splitting, which provides a better experience in this high-frequency homework process every day in school, thereby greatly improving the processing efficiency of electronic homework.
[0168] Either the answers to multiple questions are written together without distinguishing them, but the answer boxes are not divided. In the later stage, it is impossible to automatically match a certain part of the answer to a certain question, and it is impossible to conduct personalized learning statistics for students. For example, it is impossible to count the score rate of a student's trigonometric function knowledge point; it is not intelligent for teachers to use it. For example, when reviewing a test paper, the teacher hopes to automatically count the frequently wrong questions, and it is best to see the original track of the wrong answers. However, in this solution, because the answers cannot be matched with the questions, the scores cannot be automatically counted, resulting in the inability to automatically count frequently wrong questions. When the teacher explains the Zth question, the original track of the error of the question cannot be extracted and presented to the teacher; however, in the disclosed solution, the answer track in the answer area corresponds to the question one by one from the beginning, so the score can be automatically calculated, and frequently wrong questions (scores in reverse order) can be automatically generated. When reviewing, the teacher can select a question to view the original track of the student who answered incorrectly.
[0169] Or the answers to multiple questions are written together, and the answers to which questions need to be uploaded; at this time, the segmentation of students' answers depends on the clear segmentation when students take photos, but in actual use, students often write several questions on a piece of paper. If the photos are segmented, the photos need to be cropped into rectangular photos with various aspect ratios. These photos of different sizes and aspect ratios will be scaled after being displayed on the teacher's side, and the teacher will have to read them with different font sizes, which is a poor experience; this situation does not exist in this solution. The font size of the answers before and after is unified, which is convenient for teachers to correct, thereby effectively improving the user experience.
[0170] In addition, the existing technical solutions cannot automatically correspond the answer area to the question, and it is even more impossible to automatically correspond the sub-questions and their answer areas; while in this solution, each answer area can be automatically corresponded to each sub-question, ensuring the automation, standardization and rationalization of the electronic homework processing process.
[0171] In one feasible solution, the topics to be processed include several topic levels, and each topic level corresponds to several sub-topics;
[0172] Step S104 includes:
[0173] When it is determined that the sub-question belongs to the handwriting question type, a handwriting answer area of the sub-question is generated on the first preset blank canvas to obtain multiple handwriting answer areas corresponding to the question to be processed.
[0174] Among them, the pending questions generally correspond to one or more question levels, and each question level generally corresponds to one or more sub-topics. The number of question levels of different pending questions and the number of sub-topics under each question level can be the same or different, and are determined or adjusted according to the actual question-setting requirements.
[0175] Each large question includes multiple pending questions, and each pending question is composed of one or more sub-questions. Specifically, it is based on: judging whether there is a pre-set type of question number in the document, such as "(1).", "(2).", "3"...., if so, extracting and marking it. Through the hierarchical relationship between the pre-set question numbers, "1.", "2.", "3." corresponds to the second level, "(1), (2), (3)" corresponds to the third level, (11), (12), (13)..., (21), (22), (23)..., (31), (32), (33)... corresponds to the fourth level, and so on. Therefore, for large questions of any number of levels, it is possible to comprehensively, accurately and efficiently extract the questions at any level, and judge whether the corresponding questions belong to handwritten questions; wherein, each sub-question is identified by extracting the question number automatically, in batches or one by one, and whether it belongs to a handwritten question is automatically judged based on the judgment scheme of the above scheme question type. In addition, a sub-question can have n answer areas, which is convenient for teachers to follow in their grading habits. They can first mark the answer area where the answer is located, rather than looking for the answer in the entire answer area.
[0176] Specifically, as shown in Figures 6 and 7, Figure 6 shows an identified question containing three handwritten sub-questions a, b, and c, and Figures 7 and 8 correspond to the handwritten answer areas of these handwritten questions.
[0177] In this solution, for a question to be processed, all sub-questions under all question levels are identified and processed in order of question number, or multiple sub-questions are identified and processed at the same time. As long as one or more sub-questions are handwritten questions, corresponding answer areas are automatically generated for each sub-question. That is, this solution can realize batch recognition and processing of multiple sub-questions, and at the same time, it can batch generate matching answer areas for each handwritten question identified, ensuring the electronic and intelligent generation efficiency of the answer areas for handwritten questions in the questions;
[0178] In addition, such a hierarchical division of questions can also facilitate later scoring, learning situation analysis and statistics, etc. For example, it can be broken down into sub-questions. For example, multiple sub-questions in the same question test different knowledge points. Through the scores of students in the answer area of each sub-question, it can be inferred which knowledge point the students have mastered better and which knowledge point they have problems with, so as to facilitate targeted tutoring of students to improve the overall teaching quality.
[0179] In addition, since the existing solutions cannot automatically correspond the answer area to the question, it is even more impossible to automatically correspond the sub-questions and their answer areas. However, each answer area in this solution can automatically correspond to each sub-question, thereby further ensuring the automation, standardization and rationalization of the electronic homework processing process.
[0180] In one feasible solution, step S104 includes:
[0181] When there are multiple sub-questions that belong to the handwriting question type, the handwriting answer area of all the multiple sub-questions that belong to the handwriting question type is generated simultaneously;
[0182] The handwriting answer area of each sub-question is displayed following the location of the corresponding sub-question; or, the handwriting answer areas of multiple sub-questions are displayed sequentially and centrally based on the order of the questions.
[0183] In the existing solutions, either the question text and the answer content are on two separate sheets of paper, and after printing, the position relationship cannot be flexibly selected according to needs, which is not flexible enough and not convenient for overall management of the assignment; or the division of the answer area into sub-questions is not supported, and at this time there is no content related to the answer area of the sub-questions.
[0184] In this solution, for multiple handwritten questions in the processing questions, all the matching answer areas are generated synchronously to ensure the efficiency of electronic processing.
[0185] The display mode of multiple answer areas includes a first display mode: automatically generating a matching answer area directly below all sub-topics belonging to the handwritten question; or, a second display mode: these multiple answer areas can be arranged and displayed in order from top to bottom below the entire pending question in the order of question numbers. The user can operate the third specific button corresponding to the pending question to switch the display mode by clicking; wherein, the first display mode can be used by default. That is, the flexibility of the display mode is achieved, the different usage and operation requirements of different users are met, and the user's usage experience is improved. Specifically, as shown in Figures 6-8, it is identified that a certain question contains three handwritten sub-topics a, b, and c. Figure 6 shows the stem content of these three handwritten questions. Based on the above acquisition method, the handwritten answer area shown in Figure 7 can be obtained, that is, each answer area is displayed directly below the corresponding handwritten question; or it can be displayed in the manner shown in Figure 8, that is, the three handwritten answer areas corresponding to the three handwritten questions are displayed together.
[0186] The display method of the answer area can be determined by the following scheme:
[0187] Solution (1): Use the PDF of the original test paper to obtain an image, use optical character recognition technology to obtain the relative position of the question stem and other parts in the test paper, and obtain the distance between two adjacent sub-question stems, and judge whether the distance is greater than the set threshold (2 cm or the height of two lines of text). If not, it is determined that these sub-questions are put together, and the answer area needs to be generated after these questions, and a control instruction is generated. At this time, the handwritten answer area corresponding to each sub-question is displayed sequentially and centrally based on the order of the questions; if so, it is determined that the answer space needs to be increased in the two adjacent question stems, and it is determined that the corresponding answer area needs to be generated immediately after each sub-question, and a control instruction is generated to control the display of the corresponding sub-question position;
[0188] Solution (2): After accumulating a certain amount of question arrangement pattern data (i.e., question + arrangement pattern data pairs), extract the following features from the questions: question attribute features, including subject, grade, exam source, etc. (directly obtained based on question label information); high-dimensional features of question information, i.e., extracted through a pre-trained paragraph semantic model (e.g., Sentence BERT); add the question attribute feature vector and the high-dimensional features (or feature fusion) to obtain the question feature vector, and then use a similarity algorithm to calculate the vector cosine similarity of the question feature vector with all the questions in the accumulated data (e.g., Euclidean distance, Manhattan distance, K-means algorithm, etc.), calculate the question with the closest similarity, and use its corresponding question display mode as the regional display mode of the current question;
[0189] Solution (3): For all the sub-topics of handwritten questions in all previous electronic homework, a preset classification algorithm is used for classification processing (such as decision tree, support vector machine (SVM), random forest and gradient boosting decision tree (GBDT), etc.); wherein, feature extraction, basic information of the question including question length, subject, etc., is extracted from the basic information of the question using the BERT model to obtain text semantic features, and all features are spliced to obtain an n-dimensional feature vector; label acquisition: the size of the answer area of each question is measured and divided into n categories, each category represents the range of the answer area size, and the SVM classifier is trained; a correspondence between the question type and the corresponding answer area is established, and for any actual sub-question, the corresponding question type is obtained using the SVM classifier, and then the regional display mode of the sub-question is automatically generated according to the pre-established correspondence;
[0190] Solution (4): Implementation based on the algorithm model: take each sub-question as input and the corresponding regional display mode as output, and use a preset network model (such as an SVR regressor model, using mean square error loss) to perform model training to obtain a regional display mode determination model for determining the size of the answer area; the specific model training process is similar to solution (3) and will not be repeated here; for any actual sub-question, input the corresponding sub-question into the regional display mode determination model to output the corresponding regional display mode; of course, as new handwritten questions are updated, the regional display mode determination model is regularly updated to ensure the accuracy of regional display mode determination;
[0191] Solution (5): Based on the subject and the question types and settings of the previous year's test paper, the system automatically selects a recommended area display mode.
[0192] It should be noted that in the actual model training process, the judgment of whether it is a handwritten question to generate the answer area, the area display size, and the area display mode can be implemented in the same algorithm model, that is, determined at the same time to ensure the efficiency of model training; in addition, it can also be trained independently according to actual needs, or trained on data of any two dimensions; the specific training method will not be elaborated here.
[0193] In addition, the answer display area may have the following situations, which are compatible with multiple settings:
[0194] (1) The same question or sub-question can have multiple answer areas, as shown in Figure 6a. In this question, the average score of the test corresponds to one answer area, and the median score of the test corresponds to another answer area.
[0195] (2) The same question or sub-question can have multiple answer areas, and the answer area can be suspended at any desired location; as shown in Figure 6b, the horizontal line position in the question stem and any area below the question stem can be the answer area;
[0196] (3) The answer area can be in the middle of the question stem or below the question stem, as shown in Figure 6a and Figure 6e;
[0197] The answer area can also be above the question stem: if the question stem contains content indicating that the question should be completed in the front (such as "Please fill in the answer in the area above"), an answer area of a set size or a sufficiently large size will be automatically generated in the blank area in front of the question stem; or, based on manual triggering, an answer area can be generated in front of the question stem, as shown in Figure 6c;
[0198] (4) The same question or sub-question can correspond to a combination of a floating answer area and an answer area, as shown in Figure 6d;
[0199] (5) The answer area can have an underlying picture, and students continue to answer on the underlying picture, as shown in Figure 6e;
[0200] (6) The floating answer area in the table, as shown in Figure 6f, can be set according to the specific circumstances of the actual question; among them, different settings can be flexibly combined to meet the examination requirements of different subjects and different questions.
[0201] In one feasible solution, as shown in FIG9 , after step S104, the following steps are further included:
[0202] S105, receiving a region size adjustment instruction;
[0203] S106. Adjust the size of the handwriting answer area based on the area size adjustment instruction.
[0204] Among them, by obtaining the coordinates of the current pen position and the current answer area size, the system will automatically adjust the size of the answer area; when it detects that the user writes a track within a certain distance close to the current bottom edge, the area size will be automatically increased without the user having to click and adjust it multiple times, which ensures the intelligence and rationality of the operation and improves the user's operating experience.
[0205] In this solution, when the user operates in the answer area of a handwritten question, if the default generated answer area is considered to be too small (such as not enough space to write the answer, the answer area cannot accommodate corrections or notes, etc.) or too large, the size of the answer area can be flexibly adjusted according to actual needs to meet the different area ranges required by different students and teachers when using the answer area, avoiding some disadvantages of the answer area of handwritten questions in existing solutions, such as either there is no restriction on where the answer can be written, and students can write wherever they want, resulting in the inability to automatically match the student's answer content with the question they are answering; or the answer area size cannot be adjusted. If the student cannot write and writes outside the answer area, it cannot be scanned into the system later, resulting in the inability to enter all answers and no score.
[0206] In one feasible solution, the default size of the handwritten answer area is positively correlated with the score of the corresponding question to be processed.
[0207] The answer area that matches the handwritten question is: for example, the answer box size matches the content of the reference answer of the sub-question, etc., thereby effectively ensuring the rationality of the presentation of the answer area, and as much as possible, there is no need for other manual operations to adjust the answer area, thereby ensuring the user experience.
[0208] In this solution, the size of each answer area is not randomly generated, nor is it generated using a uniform size, nor is it determined manually, but is determined based on the score of the corresponding question; generally, questions with higher scores have corresponding answer content, and the required answer area size is also larger; conversely, questions with lower scores have corresponding answer content, and the required answer area size is also smaller, thereby effectively ensuring the rationality of the presentation of the answer area, making it unnecessary to perform other manual operations to adjust the answer area as much as possible, reducing subsequent size adjustment operations, and thus ensuring the user experience.
[0209] In one feasible solution, a first size adjustment plug-in is correspondingly provided at a first preset position of the handwriting answer area;
[0210] Step S105 includes:
[0211] In response to the first size adjustment plug-in, obtaining a region size expansion instruction;
[0212] Step S106 includes:
[0213] Increasing the size of the handwritten answer area to a first preset size based on the area size expansion instruction;
[0214] A second size adjustment plug-in is correspondingly provided at a second preset position of the handwriting answer area;
[0215] Step S105 includes:
[0216] In response to the second size adjustment plug-in, obtaining a region size reduction instruction;
[0217] Step S106 includes:
[0218] reducing the size of the handwritten answer area to a second preset size based on the area size reduction instruction;
[0219] In this solution, the area range of the answer area can be automatically extended according to the user clicking the area border or a specific button. Each time the corresponding specific button "+" is clicked, it will automatically extend outward along the position of the border by the set step length. In this way, the answer box area can be flexibly and personalized adjusted at any time and quickly; for example, when a student clicks the button "+" below the answer area, it will automatically extend downward in the vertical direction for a fixed step length (such as 1cm, or other values). Each click will continue to extend downward until the set number of extensions is reached and the extension will stop (for example: the upper limit of the extension number is pre-set to 15 times); if the lower edge of the currently extended answer area exceeds the current display interface, the scroll bar on the side of the answer area will be enabled, and the user can display the entire answer area by operating the scroll bar; or, when it is extended to the bottom of the current page, it will automatically extend to the next blank page, so that students can continue to extend to meet the personalized answer area adjustment needs of different students as much as possible, which is convenient for the subsequent direct printing of the answer content into a book, etc.
[0220] Specifically, referring to Figures 10-12, click the "+" button below the answer area, and Figures 10, 11, and 12 will extend the corresponding answer areas downward in turn; conversely, click the "-" button below the answer area to reduce the corresponding answer areas upward in turn.
[0221] Specifically, although the scope of the answer area can be extended, it cannot be extended indefinitely. This allows it to meet the needs of students and teachers while avoiding the waste of answer space due to students' bad answering habits, and also achieves the effect of standardizing students' answering behavior.
[0222] If after automatically extending outward, the user finds that the area has expanded too much and needs to restore some areas, he or she can click the "-" button below the answer area to automatically reduce the fixed step size (such as 1cm, or other values) in the vertical direction. Each click can continue to reduce it until the desired answer area is reached, thereby further achieving a more flexible adjustment effect, and to a great extent meeting the user's actual answering scenario usage needs.
[0223] In addition to the above methods for adjusting the answer area, other methods can also be used. For example, the user can expand the answer area downward by clicking on the lower edge of the answer area, and can shrink the answer area upward by double-clicking on the lower edge of the answer area. The specific method can be designed and adjusted according to the needs of the scenario.
[0224] In one feasible solution, a third size adjustment plug-in is correspondingly provided at a third preset position of the handwritten answer area;
[0225] Step S105 includes:
[0226] In response to the third size adjustment plug-in, a parameter input interface is displayed;
[0227] Receive external input area size parameters based on the parameter input interface;
[0228] Step S106 includes:
[0229] Adjusting the size of the handwritten answer area to a third preset size based on the area size parameter;
[0230] In this solution, a dedicated button is pre-set at the set position of the answer area. After clicking the button, a size area parameter setting pop-up window will be displayed in the system interface. The user can enter the size range area parameters of the answer box area in the pop-up interface, and click Finish to complete the area range adjustment of the answer area. In this way, a more refined and accurate size adjustment of the answer area can be achieved, thereby meeting the needs of more demanding users and further improving the user experience.
[0231] When the user's handwriting leaves the current answer area, the actual writing area height is automatically calculated based on the position of the track on the canvas, and the canvas height is adjusted. The excess blank area below can also be automatically cut off. Specifically, the position data corresponding to the latest writing track in the answer area and the position data of the lower border of the current answer area are obtained, and the distance between the two positions is calculated in real time. When the distance is less than the set value, the lower border is automatically extended downward by the set distance. If the user erases the written content during this process, making the distance greater than or equal to the set value, the lower border is automatically retracted by the set distance. The retracted answer area is no smaller than the size of the initially generated answer area.
[0232] In one feasible solution, the handwritten answer area supports obtaining writing trajectory-related data;
[0233] Wherein, the writing track associated data is stored in a preset storage space;
[0234] The writing trajectory associated data includes a plurality of writing trajectory data segments, and the first time frame data and / or the answer writing speed data corresponding to each writing trajectory data segment;
[0235] On the student side, the writing trajectory data is the answer trajectory data;
[0236] On the teacher side, the writing track data is the answer correction data.
[0237] The preset storage space includes local storage and / or cloud storage, which makes it convenient to directly obtain the corresponding writing trajectory associated data in subsequent specific queries or call requirements; the writing trajectory associated data can be transmitted to the preset storage space in one or more of the following ways: starting to write to activate the current answer area, leaving the current answer area, and staying long enough (such as 30 seconds).
[0238] On the student side, the writing trajectory data is the answer trajectory data;
[0239] On the teacher side, the writing track data is the answer correction data.
[0240] Among them, the collection of writing trajectory-related data in the answer area can be carried out at as high a frequency as possible (such as collecting once every 0.1 seconds) without affecting the trajectory restoration process, so as to obtain sufficient data and thus ensure the accuracy and reliability of subsequent data analysis.
[0241] The image of the student's answer content is identified, and the identified content is compared with multiple scoring points of the question to calculate whether each scoring point is reflected in the answer content; if so, determine in which area of the answer content image the content corresponding to each scoring point is located, and retrieve the start time and duration of each handwriting in this area to calculate at which time point each step of the student starts and ends, and obtain the timeline information of the student when doing the question.
[0242] That is, obtain the student's writing content corresponding to each scoring point of the current question, then find the corresponding trajectory data in the entire trajectory, obtain the start time and end time when the student writes the content of each scoring point, and calculate the corresponding time spent; it can also be known at the same time that the other time is the time for writing incorrect answers and other dimensions; in addition, according to needs, a chart corresponding to the timeline information can be formed for a more intuitive understanding.
[0243] Specifically, the following time periods can be obtained based on the above method:
[0244] Question review time: the time from entering the question page to the first stroke of the pen;
[0245] Time taken for the first point: Based on the answer score, the time taken by the student to obtain the score is obtained from the background data;
[0246] Second score duration: Based on the answer score, the time it takes for students to obtain the 2 points is obtained from the background data;
[0247] And so on, we can get the time it takes for the student to obtain each point;
[0248] The time taken by the student and similar students in this step can be statistically distributed for horizontal comparison; for example, the ranking of the student's time in the overall normal distribution (such as the 5% with the longest time) can be obtained, or the difference between the student's time and the average time can be obtained, if it is greater than n standard deviations.
[0249] After obtaining the normal distribution graph, if a student's first step ranks 30% longer, the second step 40% longer, and the third step 10% longer, then the student's weakest link is identified as the teaching content involved in the third step, which is the area where the student has the greatest potential for improvement. This allows us to focus on the specific step or steps where the student is stuck or where their grasp of knowledge is weak, accurately identifying the student's bottlenecks and weaknesses for targeted improvement. Of course, for students who are slow at answering questions, we can also analyze whether the problem is slow to review the questions or slow to calculate, providing a very specific understanding of each student's situation for targeted intervention and tutoring.
[0250] Among them, the automatic recognition of the scoring points of the student's answer content can be achieved in the following ways:
[0251] (1) The scoring points of students' answers are determined through the scoring point output model. Specifically, several answers to the same question are used as input, and each scorable content in each answer is marked and used as the output of the model. The model is trained using a preset network model to obtain a scoring point output model. For any answer to a question, it is input into the scoring point output model to automatically output multiple scorable contents (scoring points).
[0252] (2) The teacher marks each scoreable point (manually marks the correct area), and then uses OCR recognition technology to automatically identify the corresponding scoreable content;
[0253] (3) The student's answer content is identified through OCR recognition technology, and then the answer content is compared with the reference answer content to obtain the number of places where the overlap between the two is greater than the set value, so as to automatically identify the corresponding content that can be scored. In this solution, the answer area supports full, real-time, and dynamic tracking of all operations of students and teachers in the answer area, as well as the time corresponding to each writing track; for the student side, the student's current answer track is recorded and processed in real time. By analyzing the answer track over time, it is possible to analyze which steps of the student's current question take a long time, so as to better understand the student's weaknesses and help improve the student's mastery level in a targeted manner, thereby improving the learning situation in the entire teaching scenario, that is, to achieve timely and accurate positioning of the student's problems, without the need for manual positioning of track segments, and greatly improving processing efficiency. Specifically, for example: tracking the time when the student first started writing after entering the current handwriting question, the time when the writing was completed, the total time spent on answering the question (that is, the sum of the time of several answer tracks), the answer speed, the written answer content, and the withdrawal and modification operations during the writing process.
[0254] For the teacher side, the real-time record of the teacher's correction track for the student's answer content in the current answer area is convenient for other supervisory sides to trace back the teacher's correction data and analyze the seriousness of each teacher's correction of homework, etc., so that the teacher's correction data can be comprehensively, reasonably and effectively supervised. At the same time, it is also convenient to make targeted improvement suggestions for each teacher, thereby further optimizing the overall teaching quality.
[0255] In one feasible solution, the answer box area supports dynamic following display of writing trajectory-related data in the target end.
[0256] For example, in class, the teacher assigns an example question on the blackboard and the students write below. After the live broadcast function is turned on, the teacher can see each student's answer in real time on his or her terminal, without having to run to each student to check the details, thus greatly simplifying the supervision and management process.
[0257] Among them, the page is refreshed regularly according to the preset sampling frequency, and the pictures corresponding to the writing trajectory related data are obtained in time and reported to the teacher's end, so that the teacher can view it in time; in the actual homework scenario, different sampling frequencies can be used according to the different types of homework (classroom homework, after-school homework), and the data transmission frequency is differentiated. The data collection frequency of important homework is higher, and the data collection frequency of other non-important data is lower, so as to ensure the overall data management efficiency and rationality.
[0258] In this solution, all writing trajectory data in the answer area can be dynamically followed and displayed, that is, real-time broadcast of user operations in the display area can be achieved; different target ends have different trajectory data viewing permissions, including but not limited to: the teacher end can track and view the real-time writing trajectory data of each student in the student end, and the supervision end can track and view the real-time correction trajectory data of each teacher in the teacher end, etc.
[0259] Specifically, for example, in an invigilation scenario, a teacher can open an electronic homework being written by any student at the teacher's endpoint, view the current view of the student's answer interface, and view in real time where the student is currently doing the questions, what content he or she has written, etc.; for the supervision end (such as the school end), you can click on the homework being corrected by any teacher and view the current view of the teacher's correction interface, so as to achieve the purpose of spot checking and supervising the teacher's correction work, thereby further ensuring the optimization and improvement of the overall teaching quality.
[0260] In one feasible solution, the handwriting answer area supports displaying the operation layer of the student terminal in the first preset user terminal;
[0261] The handwritten answer area supports playback of answer track data stored in the preset storage space in the second preset user terminal;
[0262] The handwritten answer area supports displaying the teacher's operation layer in the third preset user terminal;
[0263] The handwritten answer area supports playback of answer and correction data stored in the preset storage space in the fourth preset user terminal;
[0264] Wherein, the corresponding data playback mode is determined based on different selection instructions.
[0265] Specifically, the first preset user terminal, the second preset user terminal, the third preset user terminal, and the fourth preset user terminal are operating terminals with operating permissions for the teacher terminal, student terminal, etc. The specific corresponding operating permissions are determined by pre-settings, and of course can be readjusted according to actual needs.
[0266] In this solution, the operation layer of the student side can be viewed on the teacher side, and the student's answer trajectory data in each answer area can be replayed on the teacher side. The student side can also review its own answer trajectory data in each answer area; the operation layer of the teacher side can be viewed on the supervision side (such as the school side), and the teacher's answer correction data in each answer area can be replayed on the supervision side. The student side and the teacher side can also replay the teacher's answer correction data in each answer area, etc.; the operation layer of the operation side that has viewing permission and the stored answer trajectory data can be replayed in any preset user side to achieve the effect that the corresponding answer situation can be supervised; in addition, for the replayed content, the normal playback speed can be used, and the speed can be doubled according to personal viewing needs or preferences (such as 0.75 times, 1.25 times, 1.5 times, etc.); or some content can be skipped, and only the trajectory behavior within a certain period of time is viewed and analyzed to meet the different viewing needs of different objects and ensure the user's usage experience.
[0267] In an implementable solution, the handwriting answer area is provided with a corresponding basic functional component, and the basic functional component includes a number of basic functional plug-ins.
[0268] Specifically, as shown in Figure 13, users can use the electronic pen to circle and draw in the handwritten answer area (i.e., write traces); they can also click different locations above the handwritten answer area to access different basic function plug-ins; as shown in Figure 14, the electronic pen settings plug-in is called to allow users to select pen color, handwriting thickness, etc. Other functions are similarly operated and will not be detailed here.
[0269] In this solution, it is taken into account that students or teachers may have various different operational requirements when answering questions or correcting questions in the answer area, such as calling the electronic pen tool to select a preferred electronic pen (different users can choose according to their personal operational preferences, such as selecting the type and color of the electronic pen), calling the eraser tool for corrections, calling the draft paper tool for calculations, etc.; in this solution, while automatically forming the answer area, the corresponding operation function setting plug-in, i.e., the toolbar, is also automatically generated simultaneously, so as to achieve the purpose of helping students input by hand and helping teachers correct by hand. Compared with the teaching scenarios of paper answering or dot pens in the existing solutions, the degree of electronicization of homework is generally low. This solution achieves the goal of meeting the user's several basic answering operation requirements while also meeting the user's personalized usage needs, making the user operation more flexible and further improving the user's usage experience.
[0270] In an implementable solution, the basic function plug-in includes a draft paper calling plug-in, which is used to call or hide the draft paper;
[0271] The draft paper area where the draft paper is located is generated along with the corresponding handwritten answer area, and each handwritten answer area corresponds to a draft paper; or, all handwritten answer areas under each pending question correspond to a draft paper;
[0272] The draft paper is dynamically stored in the preset storage space of the smart terminal.
[0273] In one feasible solution, the attribute information of the draft paper is adjustable;
[0274] The display style of the scratch paper area where the draft paper is located is different from the display style of the handwritten answer area;
[0275] When calling the scratch paper, determine whether the eraser function of the handwritten answer area is disabled based on the actual selection operation;
[0276] Among them, the attribute information of the draft paper includes but is not limited to transparency, size, color, draggable position, etc. It can be flexibly adjusted according to the attributes of the draft paper to meet the usage habits of different users, further ensuring the user's answering experience.
[0277] When the draft area is called, it will be displayed floating on the right side of the current answer area. Users can drag it to any position in the current answer interface as needed.
[0278] The display style needs to ensure that the current handwritten question content and answer content can still be seen when using the draft paper, such as increasing the transparency of the draft paper;
[0279] When users view the draft area and the answer area, they can use the font color to distinguish which content belongs to the draft area and which content belongs to the answer area; for example, the answer area is generally set to a white background, and the font color of the written content is black; while the draft area is generally set to a black or gray background, and the font color of the written content is white.
[0280] When calling the scratch paper, determine whether the eraser function of the answer area is disabled based on the selection operation.
[0281] In this solution, for example, for a simple physics question, students generally need to use draft paper to perform calculations and deductions. The draft paper can often reflect the student's intermediate thinking process and processing process of the question. Considering that when the teacher corrects the student's answer and finds an error, he can go to the draft area to check the student's thinking and find out the specific reason for the student's error. For example, a student made a writing error in a certain step of the calculation, which led to an error in the subsequent calculation result, but the overall thinking logic process was completely correct. In this way, the student can be corrected in a targeted manner instead of assuming that the student does not understand all the knowledge points of the physics short answer question, thereby helping to achieve better teaching results.
[0282] At the same time, when students find that their answers are wrong, they can call up the draft paper again and check whether there are any errors in the draft. This ensures that the corresponding draft content can be found quickly, which improves the students' operating experience.
[0283] In addition, the process of students writing homework is completely built into the electronic homework system disclosed in the present invention. Students do not need to open a drawing board software outside the system to make drafts, which makes it easier to call or hide the draft function more quickly and flexibly, allowing students to complete all answering operations in the electronic homework system of this solution, and making the online proctoring function possible.
[0284] In one feasible solution, the draft paper area is located on the first preset blank canvas where the handwritten answer area is located;
[0285] The draft paper area is displayed before or after the handwritten answer area;
[0286] The draft paper area and the handwritten answer area are displayed side by side in separate areas.
[0287] In this solution, the draft paper area and the answer area are located on the same canvas. The draft paper area can be displayed floating above the answer area, or it can be sunk to the bottom of the answer area for display. The specific display mode can be switched through specific operations (such as double-clicking in the front display area, etc.), thereby ensuring the flexibility and convenience of switching; in addition, the draft paper area can also be displayed side by side with the draft paper, including side by side display in the horizontal direction and side by side display in the vertical direction. Of course, it can also be displayed in other more feasible ways according to other actual needs, which will not be repeated here.
[0288] Specifically, referring to Figures 15 and 16, the draft paper area is displayed floating above the answer area. Students can flexibly adjust the transparency of the draft paper area by dragging the slider above the draft paper area (Figure 15 is more transparent, and Figure 16 is less transparent); the draft paper can be turned off by the cross in the upper right corner of the draft paper area; if the draft paper still needs to be called, click the draft paper call plug-in again to enable the draft paper function.
[0289] In one feasible solution, the draft paper area supports obtaining draft trajectory associated data;
[0290] Among them, the draft trajectory associated data is stored in a preset storage space;
[0291] The draft trajectory associated data includes a plurality of draft trajectory data segments, and second time frame data and / or draft writing speed data corresponding to each draft trajectory data segment;
[0292] In the student side, the draft track associated data is the draft writing track data;
[0293] On the teacher side, the draft track associated data is the draft correction track data.
[0294] Among them, the preset storage space includes local storage and / or cloud storage, which makes it convenient to directly obtain the corresponding draft trajectory associated data under subsequent specific query or call requirements; the draft trajectory associated data can be transmitted to the preset storage space in one or more of the following ways: starting to write to activate the current draft area, leaving the current draft area, and staying there for a long enough time (such as 30 seconds).
[0295] Specifically, the collection of draft trajectory-related data in the draft paper area can be performed as frequently as possible (e.g., once every 0.1 seconds) without affecting the trajectory restoration process, so as to obtain sufficient data and thereby ensure the accuracy and reliability of subsequent data analysis.
[0296] In this solution, the answer area supports full, real-time, and dynamic tracking of all operations of students and teachers in the draft paper area, as well as the time corresponding to each draft track. While capturing the writing track-related data in the answer area, it relies on the draft track-related data to assist teachers and students in tracing back some problems in the answering process, so that teachers can better understand students' test-taking situation and students can better understand teachers' correction situation, so as to make targeted suggestions or modifications, thereby further optimizing the overall teaching quality.
[0297] In one feasible solution, the answer box area supports dynamic tracking and display of the draft track associated data of the draft paper on the target end.
[0298] Among them, the page is refreshed regularly according to the preset sampling frequency, and the pictures corresponding to the draft trajectory related data are obtained in time and reported to the teacher's side, so that the teacher can view it in time; in the actual homework scenario, different sampling frequencies can be used according to the different types of homework (classroom homework, after-school homework), and the data transmission frequency is differentiated. The data collection frequency of important homework is higher, and the data collection frequency of other non-important data is lower, so as to ensure the overall data management efficiency and rationality.
[0299] In this solution, all draft contents in the draft paper can also be dynamically followed and displayed, that is, real-time broadcast of user operations in the draft area can be achieved; different target ends have different data viewing permissions, including but not limited to: the teacher end can track and view the draft content written in real time on the draft paper by each student on the student end, and the supervision end can track and view the real-time correction content on the draft paper by each teacher on the teacher end, etc.
[0300] Specifically, in the invigilation scenario, the teacher can open the electronic homework being written by any student at the teacher's endpoint, view the current view of the student's answer interface, and view in real time whether the student is currently writing in the answer area or the draft area, as well as the content written in the draft area in real time, so as to obtain more accurate information about each student's specific answer, facilitate the teacher's comprehensive understanding of the students' answer status and supervision purposes, and facilitate the realization of high-quality teaching requirements.
[0301] In one feasible solution, the draft paper area supports displaying the draft writing trajectory data of the student terminal in the first preset user terminal;
[0302] The draft paper area supports playback of the draft writing trajectory data stored in the preset storage space in the second preset user terminal;
[0303] The draft paper area supports displaying the teacher's draft correction track data in the third preset user terminal;
[0304] The draft paper area supports playback of the draft correction track data stored in the preset storage space in the fourth preset user terminal;
[0305] Wherein, the corresponding data playback mode is determined based on different selection instructions.
[0306] Specifically, the first preset user terminal, the second preset user terminal, the third preset user terminal, and the fourth preset user terminal are operating terminals with operating permissions for the teacher terminal, student terminal, etc. The specific corresponding operating permissions are determined by pre-settings, and of course can be readjusted according to actual needs.
[0307] In this solution, the operation layer of the student side can be viewed on the teacher side, and the draft trajectory associated data of the students in each draft area can be replayed on the teacher side. The student side can also review the draft trajectory associated data in each draft area; the draft correction data of the teacher in each draft paper area can be replayed on the supervision side (such as the school side), and the student side and the teacher side can also replay the draft correction data in each draft area, etc.; the operation layer of the operation side that has viewing permission and the stored draft trajectory data can be replayed in any preset user side to achieve the effect of supervisory of the corresponding answer situation; in addition, for the replayed content, the normal playback speed can be used, and the speed can be increased according to personal viewing needs or preferences (such as 0.75 times, 1.25 times, 1.5 times, etc.); or some content can be skipped, and only the trajectory behavior within a certain period of time is viewed and analyzed to meet the different viewing needs of different objects and ensure the user experience.
[0308] Specifically, on the teacher's side, the teacher can see the content of the draft written by the students. When the teacher is not clear about the answer content or has doubts after only looking at the answer area, he can check the draft paper function to view the content of the draft paper and find the mistakes made by the students when drafting. For example, the student's calculation logic process is correct, but the calculation on the draft paper is wrong, and the intermediate result is copied incorrectly from the draft paper. At this time, feedback can be given to the students through messages, etc., and the real reason why the students made wrong answers can be pointed out in a targeted manner; on the student side, students can see where the teacher circled the mistakes in their drafts when correcting them, so as to quickly find out the reasons for their own wrong answers; when students check the content of the teacher's corrections on the draft part, if they don't understand the teacher's corrections, they can also replay the correction process slowly and repeatedly until they understand it clearly, thereby achieving higher quality teaching results.
[0309] In one feasible solution, the obtaining method further includes:
[0310] Get the actual writing data in the dynamic follow-up display answer area;
[0311] The actual writing data includes writing track associated data, or is based on the writing track associated data and the draft track associated data;
[0312] Analyzing the actual writing data using a preset abnormal answer monitoring rule to obtain a first analysis result;
[0313] Based on the first analysis result, it is determined whether there is any abnormality in the answering behavior or the correcting behavior. If so, a first alarm message is generated and issued.
[0314] Specifically, the following methods are used to determine whether there are any abnormalities in answering or correcting questions:
[0315] Solution (1): Compare the written data of the students' answers to this question with the written data of the answers to previous questions, and determine whether there are any anomalies based on the data characteristics. Process the time per minute based on the score and the answering time, statistically analyze the time per minute data of all students, calculate the mean and standard deviation using normal distribution, mark the data beyond n times the standard deviation as potential anomalies, and then make further judgments. Count the time all students spend answering the same question, calculate the mean and standard deviation using normal distribution, mark the data beyond n times the standard deviation as potential anomalies, and then make further judgments;
[0316] Solution (2): Through manual proctoring and solution (1), we accumulate and confirm cheating situations and obtain a number of data pairs (student answer record - whether cheating). We use the answer record features, such as answer time, operation frequency, number of cut-outs, etc., as a high-dimensional vector as the model input. The model can be any neural network and output a two-dimensional vector, which represents the possibility of cheating. We use the accumulated data to calculate the model cross entropy loss, use the back propagation algorithm to update the model weights until convergence, and obtain a cheating confirmation model, thereby maximizing the probability of the model's correct prediction. In addition, we also conduct manual spot checks, update data, retrain the model regularly, and optimize the model to improve the accuracy of judgment.
[0317] The first analysis result of the actual writing data analysis is input into the cheating determination model to output information on whether cheating exists; if the output information indicates that there is suspicion of cheating, a prompt message is output that can specifically explain where the student's writing is abnormal (such as a student is suspected of cheating on a certain question) (such as directly displayed on the teacher's homework page for correcting the student), which is convenient for the teacher to check.
[0318] Solution (3): For the entire test paper, all students' answers in the same test are compared with each other, and each student's answer is also compared with the standard answer. If the similarity is too high (such as the preset threshold is 95%), it is judged that there is potential cheating. Among them, multiple-choice questions, short-answer questions with very simple answers, and questions with short and unique answers are not detected in this way.
[0319] In this solution, the user's real-time writing in the answer area is tracked and analyzed to automatically identify whether an abnormal situation has occurred. When an abnormal situation is determined, an alarm message is fed back to alert the user.
[0320] To further improve the accuracy of abnormal situation identification, the system considers the content of students' drafts in the draft paper area and the time spent on drafting, as well as the content of students' answers in the answer area and the time spent on answering. This can avoid situations where some students spend most of their time on the draft paper and then quickly copy the answers in the answer area, thereby misjudging the student as having copied homework. For example, if the analysis shows that a student's answer score for a certain question is high, but the sum of the actual writing time in the corresponding answer area and the actual writing time in the draft area is significantly shorter than the average total time for high-scoring questions, it is assumed that the student may have copied the answers. The analysis results are fed back to teachers, parents, etc., so that relevant personnel can intervene and handle the matter in a timely manner, allowing students to correct their mistakes in a timely manner. This can effectively reduce the occurrence of adverse situations such as copying answers. The entire analysis process does not require human participation and is fully automated, ensuring that the entire process of answering questions is traceable and controllable, thereby ensuring high-quality academic management.
[0321] In one feasible solution, an audio interaction plug-in is provided corresponding to the handwriting answer area;
[0322] The acquisition method also includes:
[0323] Responsive to the audio interaction plug-in, acquiring and storing externally input audio content and audio-related data;
[0324] In this solution, each answer area is automatically matched with an audio content input plug-in, so that when students answer questions in the answer area, if they have questions that need to be answered by the teacher, they can leave a message by clicking the button. Through voice communication, students can express their confusion more clearly, and it is also convenient for teachers to better understand students' questions and answer questions more efficiently and conveniently. At the same time, it is also convenient for teachers to leave messages by clicking the button when correcting, so as to quickly leave explanations and comments, verbally point out the problems, and facilitate students' understanding. It is also quick and labor-saving for teachers. While meeting the needs of answering questions, it enhances interactivity and further optimizes the overall teaching management effect.
[0325] The acquisition method also includes:
[0326] In response to the audio interaction plug-in, obtain and store externally input audio content and audio-related data, and synchronize the audio content and writing trajectory-related data belonging to the same time frame based on the audio-related data to merge the audio content and the writing trajectory-related data to obtain merged data;
[0327] Among them, the handwritten answer area supports the playback of merged data in the handwritten answer area on the target end.
[0328] In this solution, the audio data and writing trajectory correlation data in the answer area are processed temporally to obtain the voice, video and writing content in the same time frame; on the teacher's side, the questions in the student's voice message and the written content are matched, so that it can be clearly known when the student left the message and what content was written before and after the message, so as to provide better guidance and answer doubts; on the student's side, the teacher's notes when explaining the voice message to the student, for example, when verbally mentioning a certain error, the student's writing step is circled at the same time, forming the effect of actual face-to-face correction of homework in front of the student and the student listening to the explanation at the same time, and the teacher can review and study repeatedly many times, so as to achieve more efficient and accurate problem pointing out, answering questions and solving doubts, and then realize the full process tracking and analysis of any handwritten question in the entire electronic homework of any student, to ensure high-quality guidance for each electronic homework, so that long-term high-quality tutoring teaching can be guaranteed.
[0329] Example 3
[0330] The method for processing electronic homework in this embodiment is implemented according to the method for obtaining the answer area in embodiment 1 or 2.
[0331] As shown in FIG17 , the electronic homework processing method of this embodiment includes:
[0332] S201. Based on the preset learning objectives, select several questions from the preset question bank;
[0333] Among them, teachers can manually select questions that meet the requirements from the preset question bank, or they can select them through automatic identification and selection. Of course, in order to ensure the quality of the questions, teachers can perform quality screening on the selected questions. The specific selection method can be determined according to the actual situation.
[0334] The process of selecting a topic can be achieved in the following ways:
[0335] (1) Based on a previous test paper selected by the teacher, the text and images in the test paper are analyzed using a feature extraction model to obtain the feature data corresponding to the test paper; questions with a correlation degree greater than a set value with the feature data are matched in the entire question bank, and questions are set based on these matched questions; (2) For knowledge points that have been tagged, questions that other teachers have picked for the knowledge point are found, and questions are set directly based on these questions;
[0336] (2) Check the knowledge points to get all the related questions in the question bank, and use them as a benchmark to further select the questions you want;
[0337] (3) Determine the type of each question, manually select two questions that meet the question type from the question bank, and use this as a benchmark to automatically generate the remaining number of questions under the question type.
[0338] S202, using an acquisition method to generate a handwritten answer area corresponding to each question of the handwritten question type;
[0339] S203, generating set answer positions corresponding to other remaining questions of the non-handwritten question type;
[0340] S204: Based on the question and the corresponding handwritten answer area, set the answer position and obtain the electronic homework.
[0341] Among them, the electronic homework includes questions that support multiple media (such as voice interactive questions, video interactive questions, etc.); among them, the display of the question stem and the entry of the corresponding answer method are all achieved through multiple media channels.
[0342] Specifically, for a student A, based on the student's basic information (grade, etc.), historical performance, preset goals and other characteristics, the student B who is most similar to student A is determined in the accumulated data (student goals - assigned homework), and questions are randomly selected from student B's homework.
[0343] Among them, the similarity calculation method for determining the most similar students is: if two students have a common feature, the corresponding similarity score is recorded, and the student with the highest similarity score with student A among all students is the most similar student; it should be noted that: if all similarities are lower than the minimum similarity score threshold (such as 60% of the highest score), it is judged that there are no similar students and no recommendation is made in this way; if the similarity is the same with multiple students in the database, one student is randomly selected from the homework of all these students as the most similar student.
[0344] In addition, manual screening results can be recorded, including the features used in automatic question identification and whether the final question was adopted. When randomly selecting from eligible questions with the same features, all questions have the same probability of being selected initially. Subsequently, the probability of a question being recommended decreases based on the number of times it has been discarded. In this solution, the pre-set question bank includes a large number of questions, each of which is matched with corresponding knowledge points and difficulty levels. Different learning objectives require students to be assessed on different knowledge points, so matching questions are selected from the question bank to form electronic homework that meets the learning objectives for students to practice and master the knowledge points to achieve the corresponding learning objectives.
[0345] Once the selected question is a handwritten question, the answer area corresponding to the handwritten question is generated based on the acquisition method in the above embodiment; among them, multiple questions can be selected in batches, for example, 20 questions can be selected at one time, 8 of them can be automatically identified as handwritten questions, and corresponding answer areas are generated for these 8 handwritten questions at the same time, so as to achieve the purpose of efficiently generating electronic homework; or, according to actual needs, questions can be selected one by one, and whether the questions are handwritten questions can be identified one by one, and then the answer area for each handwritten question can be generated one by one; in addition, for the remaining non-handwritten questions (such as multiple-choice questions, judgment questions, etc.), a unified method is adopted to generate answer positions following the corresponding questions, such as check boxes, answer lines, etc.; no human intervention is required throughout the process, and handwritten questions and non-handwritten questions are automatically distinguished and processed separately to ensure that the overall generation of electronic homework is reasonable, efficient and of high quality, which can well meet the usage requirements of high-demand teaching scenarios.
[0346] Specifically, for any handwritten question selected that meets the learning objectives, a corresponding handwritten answer area will be automatically generated. These handwritten answer areas enable the creation of a single electronic assignment containing the handwritten question, which in turn enables the creation of a collection of electronic assignments containing handwritten questions, namely an electronic workbook. The implementation of electronic workbooks enables the subversive realization and optimization of the following educational scenarios: improving teacher efficiency and increasing the efficiency of the homework process. For example, through electronic homework, the system can help calculate individual student learning progress, thereby facilitating teachers to provide targeted tutoring. Students can also better achieve their learning goals through electronic homework, such as automatically collecting wrong questions and pushing exercises. It also provides information support for the organization, supervision, management and teaching of schools and teachers, such as checking the amount of homework assigned.
[0347] Compared with the solutions in existing teaching scenarios, either paper answer boxes need to be printed in advance, and students need to scan the answers after handwriting, etc., which leads to low efficiency and automation of the entire electronic homework; or special dot paper, dot electronic pen and Bluetooth module need to be used, and the computer needs to run software that receives Bluetooth signals, resulting in the inability of students to correspond one-to-one to each specific question when answering questions. Various information is isolated from each other and basically unrelated, which also makes it impossible to complete personalized learning statistics; or students need to take photos and upload them after answering the questions on paper, but students resist the additional photo and upload operation, and the font size of the uploaded pictures is different, which makes it difficult for teachers to correct them, resulting in a poor user experience for both students and teachers, resulting in low efficiency and automation of the entire electronic homework, which cannot meet high-demand teaching scenarios.
[0348] Example 4
[0349] The electronic homework processing method of this embodiment is a further improvement of the solution of Example 3. Specifically:
[0350] In an implementable solution, the handwriting answering area supports obtaining writing trajectory associated data, the writing trajectory associated data including a plurality of writing trajectory data segments and the first time frame data corresponding to each writing trajectory data segment;
[0351] As shown in FIG18 , after step S204, the following steps are further included:
[0352] S205: Obtaining the dwell time data of the teacher and / or student terminal on each question in the electronic homework, as well as the first time frame data of the handwritten answer area;
[0353] S206, analyzing and processing the residence time data and the first time frame data to obtain a second analysis result;
[0354] S207: Evaluate the grading and / or answering status of the teacher and / or student based on the second analysis result.
[0355] By recording the writing tracks at each collection moment in the answering area and its corresponding occurrence time, it is convenient to view the total time students spend answering each question and teachers spend correcting each question, including the time of single entry and exit, the time of repeated entry and the corresponding time of each exit, etc., so as to facilitate retrospective evaluation and analysis of each student's answering habits, answering ability and other information; and to evaluate each teacher's correction habits, the degree of correction care and other information.
[0356] For example, the student's question review time t1 (t1=T2-T1) is determined based on the time T1 when the student enters the question page and the time T2 when the student first touches the answer area to start answering. In this way, the average question review time t for the question in a class is calculated; if the analysis finds that a student's question review time t1 is significantly shorter than the average question review time t and the score is low, it can be concluded that the student's "question review habit needs to be improved"; it can also be determined based on whether there is repeated entry time whether the student has returned to the questions he has done to check or try again; of course, it is also possible to determine possible problems in other analysis dimensions of the students to further discover problems and make targeted adjustments and corrections, which will not be elaborated here.
[0357] In this solution, data recording and analysis based on the test-taking process and the grading process are implemented, so that the problems of each student and each teacher can be obtained efficiently and accurately, which facilitates the formation of more targeted teaching plans to guide teaching, thereby ensuring long-term high-quality teaching.
[0358] In one feasible solution, as shown in FIG19 , after step S204, the following steps are further included:
[0359] S208, counting the dwell time data of each entry into the same question in the electronic homework, so as to obtain a first dwell time corresponding to the same question;
[0360] S209: Obtain the sum of the first dwell time corresponding to each question in the electronic homework, so as to obtain the second dwell time corresponding to all questions;
[0361] S2010: Perform analysis based on the first stay duration and the second stay duration to obtain a third analysis result.
[0362] S2011. Determine the time consumption of grading and / or answering questions on the teacher side and / or the student side based on the third analysis result.
[0363] By recording the time a student spends on the same question each time, the total time the student spends on the same question can be accumulated; by comparing the student's score on the question with the total time spent, the student's mastery of the knowledge points tested on the question can be determined.
[0364] Specifically, if the student scores high on the question and the total time spent is similar to the average time spent on the question in the class, it is determined that the student has mastered the knowledge point tested on the question well; if the student scores high on the question but spends too much time on it (for example, more than 3 minutes longer than the average time spent on the question in the class), it means that the student's grasp of the knowledge point tested on the question is not good enough and needs further consolidation and practice. Similarly, by recording the time a teacher spends on the same question each time, the total time the teacher spends on the same question can be accumulated to obtain the total time the teacher spends on the same question; the teacher's grading efficiency can be determined by the total time spent by the teacher; the longer the total time spent, the lower the teacher's grading proficiency and the need to improve their own grading speed to improve the problem of low grading efficiency.
[0365] By counting the time spent on a single question, the total time spent by the student on the entire electronic homework can be accumulated; based on the total time spent and the total score on the entire electronic homework, an overall assessment is conducted to determine the areas for improvement of the student, and personalized tutoring and teaching are provided to different students, thereby improving the overall teaching quality.
[0366] Specifically, for the entire electronic homework, if the student's total time spent is short and the total score is high, it is preliminarily judged that the student may have plagiarized the homework, or the student's previous test scores are compared to determine whether the student has strong learning ability. If the student's total time spent is short but the total score is low, it is determined that the student is not serious about doing the questions and has not invested enough time. If the student's total time spent is long but the total score is low, it is determined that the student has difficulties and needs intervention and assistance from teachers, parents, etc. to improve. If the student's total time spent is long but the total score is high, it is determined that the student is a diligent student with no problem in mastering the knowledge, but needs guidance in thinking to improve the efficiency of doing the questions.
[0367] We collect statistics on the time it takes all teachers to grade a particular problem or electronic assignment, using normal distribution statistics to calculate the mean and standard deviation. Data outside of n times the standard deviation are marked as potential anomalies. We also collect statistics on the time it takes the same teacher to grade the same problem over time, using normal distribution statistics to calculate the mean and standard deviation. Data outside of n times the standard deviation are marked as potential anomalies. For example, if the average time it took to grade a problem in the past was 2 minutes, but it took an average of 20 seconds during the current monitoring period, this indicates that the teacher was not grading this problem seriously during the monitoring period.
[0368] The supervision module in the system accumulates the time students spend on solving problems to obtain the total time students spend on one assignment, the total time they spend studying each subject every day, and the total time they spend studying all subjects every day; similarly, weekly and monthly learning data can be obtained; based on these statistical data, the learning burden of individual students, the learning burden of the entire class, and the learning burden of the entire school can be accurately obtained, achieving the purpose of supervision and burden reduction; it can also automatically calculate and update according to the frequency to obtain the average learning time of students in each school, and derive the ranking of each school in the province and the country, and automatically mark the top-ranked schools.
[0369] In addition, a personal portrait of each student can be obtained: the percentile of the student's time investment in different subjects among similar students can be calculated. For example, if a student's mathematics study time is in the top 10% of students of the same grade in the same region, encouraging messages can be sent based on this data. Students' interests in different subjects can also be compared. For example, if the mathematics study time is in the top 10% and the study time of other subjects is at the average level, it can be inferred that the student has a stronger interest in mathematics learning, and extracurricular mathematics knowledge expansion can be recommended.
[0370] For analyzing potential anomalies: Based on the marked potential anomalies, corresponding labels can be automatically generated for each homework and fed back to teachers, students, or other relevant personnel, such as visual labels. Teachers or students can click on the label information to view the details, letting teachers know where the grading is not serious or too slow, and letting students know where the answers are not serious or too slow, playing a role in supervision, reminder, and improvement.
[0371] In this solution, by recording data such as the time teachers and students spend on each question, multiple entries into the questions, and time statistics for the entire electronic homework, we can further trace back to the problems of students' answering process and teachers' processing efficiency and learning ability in the grading process, so as to facilitate timely adjustments and improvements, thereby facilitating long-term and effective teaching management.
[0372] In one feasible solution, as shown in FIG20 , after step S204, the following steps are further included:
[0373] S2012: Counting the dwell time data of each time the user enters the same question in the electronic homework, so as to obtain a first dwell time corresponding to the same question;
[0374] S20123. Obtain first time frame data corresponding to the writing trajectory data in the handwritten answer area of the question;
[0375] S2014: Perform analysis based on the first time frame data and the first stay duration to obtain a fourth analysis result.
[0376] S2015. Determine whether there is any abnormality in the answering behavior or the correcting behavior based on the fourth analysis result. If so, generate and issue a corresponding second alarm message.
[0377] Among them, judging whether there is an abnormality in answering or correcting behavior is similar to the above-mentioned judgment method of determining whether there is an abnormality in answering or correcting behavior based on the first analysis result, so it will not be repeated here.
[0378] By recording the time a student spends on the same question each time, the total time the student spends on the same question can be accumulated; and the time information when the student writes the answer in the answer area of the question is recorded; if the total time spent by the student in processing the entire question (reading the question, thinking, writing the answer, and writing a draft) is less than a certain value, and the analysis shows that the interval between writing is very short, it is determined that the student has obviously started answering the question without thinking, and has plagiarized the homework or written the answer at random.
[0379] By recording the time a teacher spends on the same question each time, the total time the teacher spends on the same question can be accumulated; and the time information when the teacher is grading in the answer area of the question can be recorded; if the total time spent on grading is very short, but the teacher does not carefully analyze the answer process written by the students, and only ticks the correct answers, and gives 0 points if the answer is wrong, then it is determined that the teacher spent a short total time on grading, the teacher's grading speed is too fast, and the teacher is sloppy and careless in grading. Attention should be paid to timely adjusting the grading habits to ensure the quality of grading.
[0380] In this solution, by tracing back the overall time spent by each teacher and student on each question, as well as the writing trajectory data of the entire answering process and the correction process, the two are combined to analyze whether each teacher and student has any abnormal situations such as plagiarism, careless correction, etc. Once these abnormal events are detected, abnormal reminder information is automatically generated to remind relevant personnel in time (such as by publishing the alarm information to the relevant personnel's operation terminal, email, etc.), so that personnel with information receiving authority can understand and grasp the corresponding situation in time and intervene in time; at the same time, it is also convenient for teachers and students to discover their own shortcomings based on the alarm information, and make timely adjustments to avoid similar situations in the subsequent homework processing structure, so as to achieve the purpose of better teaching supervision.
[0381] In one feasible solution, as shown in FIG21 , after step S204, the following steps are further included:
[0382] S2016: Obtain the answer progress data corresponding to the question and send it to the preset storage space;
[0383] The answering progress data includes first representation data of whether the student terminal has entered the question, and answering track data of the student terminal in the answering area of the question and / or draft writing track data on the draft paper;
[0384] S2017, when the first representation data indicates that the student terminal has entered a question, controlling the first current display interface of the student terminal to be shared and displayed in the first preset user terminal;
[0385] S2018. After the student terminal submits the question, the answer track data of the student terminal and / or the draft writing track data on the draft paper are controlled to be replayed in the first preset user terminal;
[0386] Or, as shown in FIG22 , after step S204, the following steps may be further performed:
[0387] S2019. Obtain the correction progress data corresponding to the question and send it to the preset storage space;
[0388] The correction progress data includes second representation data of whether the teacher terminal has entered the question, and correction track data of the teacher terminal on the answer area of the question and / or correction track data of the draft on the draft paper;
[0389] S2020, when the second representation data represents that the teacher terminal enters the question, controlling the second current display interface of the teacher terminal to be shared and displayed on the second preset user terminal;
[0390] S2021. After the teacher completes the question correction, the correction track data of the teacher and / or the draft correction track data of the draft paper are controlled to be replayed in the second preset user terminal.
[0391] In this solution, when a student begins to enter a handwritten question, the operating terminal with monitoring authority (such as the teacher's terminal, parent's terminal, etc.) can dynamically follow and statically display all the student's answers and drafts under the question. Similarly, when a teacher begins to enter a handwritten question, the operating terminal with monitoring authority (such as the school's terminal, etc.) can dynamically follow and statically display all the teacher's answers and draft corrections under the question. In other words, the student's answers and the teacher's corrections can be displayed in real time and statically, achieving the effect of simulating on-site supervision and meeting specific teaching supervision needs.
[0392] Among them, it is possible to obtain whether any student has entered the state of doing homework or when he or she starts doing homework, so as to grasp the progress of each student's homework. In this way, it is possible to understand any student's enthusiasm for doing homework over a period of time. For example, if a student always starts to do homework after 10 o'clock every night for a week, it can be determined that the student has low enthusiasm for doing homework and is relatively slow, and timely corrections and improvements can be made.
[0393] Specifically, it mainly implements monitoring of the following four scenarios:
[0394] Scenario (1) In-class practice scenario: The teacher assigns an in-class practice and students are required to complete it in real time. The system can dynamically follow and display the answers of multiple students on the same page at the same time, so that the teacher can know the situation of each student, whether they can solve the problem, which students have faster thinking and can think of the correct method quickly, and which students have no idea temporarily or have adopted the wrong method, so as to provide targeted explanations according to the specific situation;
[0395] Scenario (2) Ordinary evening self-study scenario, the system can summarize the self-study situation of students in multiple classes and automatically mark out abnormal students. The specific marking process is: if 70% of the students in a class have started the evening self-study task, then mark out the 30% of students who have not started the evening self-study task; during the entire evening self-study process, according to the score of the questions completed by the students since the evening self-study (if the questions are not marked with scores, then calculate how many answer areas have written content), calculate the student's homework progress, and according to the distribution of all students' homework progress, get the current progress in the bottom 10% of students, and the system marks the student's progress as lagging behind; with the assistance of the system, the same teacher can manage the evening self-study discipline of multiple classes, and only needs to remind abnormal students in a targeted manner through statistical results; if the teacher needs, he can click on each student's detailed data to view the progress of the student's evening self-study homework, which question he has entered, from what time to what time he stays on which question page, from what time to what time he started writing, etc., so as to realize the viewing and supervision of specific links;
[0396] Scenario (3) Holiday homework scenario: During the entire holiday, the student's homework progress is calculated based on the score of the questions completed by the student since the start of the holiday (if the question has no score, the number of answer areas with written content is calculated); based on the distribution of all students' homework progress, the system will automatically send a message to remind the student himself, or remind parents and teachers that a student's progress is lagging behind and needs to take bold actions;
[0397] Scenario (4) Examination supervision scenario: The system will monitor and automatically record students' answering behavior, which questions they are answering from which time point to which time point, and whether they use other software during the examination. For example, the system can obtain, save, and transmit the following data: from 8:00 to 8:05, the student is using the software for answering questions in this solution; from 8:05 to 8:10, the student is not using the software for answering questions in this solution; from 8:10 to 8:15, the student is using the software for answering questions in this solution on a split screen;
[0398] During the exam, the system automatically identifies students who are not currently using the test-taking software, or who are using it in a split-screen mode, and promptly sends a reminder to the teacher to address the situation. After the exam, the system automatically calculates the total duration of each student's abnormal behavior (time not using the test-taking software + time using the test-taking software in a split-screen mode) and automatically identifies students whose abnormal behavior exceeds a certain threshold. This allows teachers to intervene and inquire in a timely manner, and can significantly reduce abnormal behavior on the student side, ensuring the feasibility and reliability of online exams.
[0399] On the teacher side, for large-scale exam marking scenarios, you can view each teacher's current grading progress; the supervisor side can also watch each teacher's grading process live, and mark which teacher's student's question was sloppy or inattentive in grading.
[0400] In one feasible solution, as shown in FIG23 , after step S204, the following steps are further included:
[0401] S2022. For multiple electronic jobs in the same batch, perform at least one of the following operations:
[0402] Obtain the completion progress of each electronic homework, dynamically track and display each electronic homework, obtain static display of each electronic homework, and obtain playback of each electronic homework;
[0403] S2023. Control the static display or playback of different handwritten answer contents corresponding to the same handwritten question in different electronic assignments.
[0404] In this solution, the handwritten answer content of the above four scenarios can be statically displayed or replayed. Specifically, the processing of electronic assignments distributed to all students in the entire class in the same batch can be tracked and supervised. For example, the completion degree of several electronic assignments corresponding to all students can be monitored, and the answer status of each student's electronic assignments can be displayed in real time. The current view of each electronic assignment can also be statically displayed, or the entire answer process of each student corresponding to each electronic assignment can be replayed, so as to realize timely and comprehensive tracking and backtracing of the learning situation of each student in the entire class, ensuring further optimization of the overall teaching management quality.
[0405] In one feasible solution, the processing method further includes:
[0406] determining a calling mode of a scratch paper in an electronic homework based on a selection instruction;
[0407] Among them, the calling mode corresponds to calling a common draft paper for use with multiple questions;
[0408] Public draft paper is reusable
[0409] In one feasible solution, independent draft paper is called for use with corresponding questions, and each question corresponds to an independent draft paper.
[0410] In this solution, all questions in the entire electronic homework can share one piece of draft paper; or each question can correspond to a piece of draft paper, so that the entire electronic homework can correspond to several pieces of draft paper; the specific type of draft paper to be called, or the simultaneous activation of two modes of draft paper, can be determined or adjusted according to actual needs to ensure the usage preferences of different users, the flexibility of operation, and enhance the user experience.
[0411] Specifically, the entire electronic homework corresponds to a piece of draft paper, which can meet the need to view archived draft papers during the formal exam; each question corresponds to a piece of draft paper, and students and teachers can find which question each part of the draft corresponds to, which is convenient for recalling and reviewing afterwards.
[0412] When grading homework, teachers can leave tracks on their draft paper. After grading, students can continue to leave tracks on their draft paper to make corrections or notes. Track data left by different users is marked with information such as the corresponding user account name and operation time. This is used to distinguish which user added different tracks and when, making it easier for different users to view them separately.
[0413] In one feasible solution, the processing method further includes:
[0414] Obtaining annotation data generated when the teacher and / or student terminal annotates the question stem area through handwriting touch;
[0415] The layer where the labeled data is located is stacked on the layer where the question stem is located;
[0416] The annotation data is stored in the preset storage space.
[0417] In this solution, students or teachers can circle and draw on the question stem, and the marked information remains on the question stem, preserving the experience of a paper test and ensuring a user experience for students or teachers. Specifically, on the student side, circling key words can help them review the question and find solution ideas. On the teacher side, circling overlooked parts can draw students' attention to certain key information in the question stem, allowing students to quickly understand where they made mistakes, thereby achieving better teaching interaction and more effective teaching goals. After the assignment is graded, students can continue to write and draw in the question stem area to make corrections or take notes. The marked content of different users can be displayed separately by user selection or default selection, such as blue for student-side marked content and red for teacher-side marked content, making it easier for students and teachers to view and understand. Of course, the display method of the marked content can also be adjusted according to actual needs. There is no specific limitation, as long as it can be distinguished and better suited to the usage preferences of different users, further ensuring a user experience for students or teachers.
[0418] In one feasible solution, the processing method further includes:
[0419] Build the topic data architecture for topic construction;
[0420] Among them, the topic data structure corresponds to several topic levels, and each topic level corresponds to several sub-topics;
[0421] Specifically, by building a question data architecture, all the components of an electronic assignment are clarified, including but not limited to the question type, each question, sub-topics at each level under the question, and corresponding identification and label data, etc., ensuring that each sub-topic can be automatically identified using logical judgment and artificial intelligence algorithms.
[0422] Obtain the subtopic identification data and subtopic label data corresponding to different subtopics, as well as the subordinate association data of the topic level to which each subtopic belongs, and store them in a storage mapping data table. In this solution, each topic in the electronic homework includes one or more topic levels, some corresponding to levels 1-2, some corresponding to levels 3, or even more levels; the number of levels can be determined or adjusted based on the question-setting requirements of different subject backgrounds, ensuring that the question settings meet the question-setting requirements while ensuring the flexibility of the question-setting operation and taking into account any required question-setting scenarios; among them, question types of various subject backgrounds can use the same set of data structure models to enter the system for electronic homework question setting.
[0423] In addition, each sub-topic under each question level has corresponding identification information (such as the identification information is 2, which identifies the second sub-topic under the current question), and label information (including but not limited to the type of knowledge point, difficulty level, and question quality). Specifically, the label data of the sub-topic makes it possible to subsequently determine the specific degree of mastery of the student based on the student's scores on different sub-topics; it also corresponds to the subordinate association data of the question level to which the sub-topic belongs (such as the subordinate association data 2.3, which means that the current question corresponds to the third sub-topic under the level below the second sub-topic under the current question. This is a three-level question data architecture), etc., and these relationship information are stored in the corresponding mapping data table to facilitate the orderly management of the overall data and facilitate data query, data modification and other operations.
[0424] In one feasible solution, the processing method further includes:
[0425] Get several selected sub-topics under the topic data structure;
[0426] Based on the external selection instruction, determine the target subtopic to be deleted, so as to update the selected new subtopics under the topic data structure;
[0427] Among them, the number of question levels is expandable.
[0428] In this solution, for several sub-topics that already exist under a question, one or more of them can be selected and deleted as needed to update the final number of sub-topics under the question to form the final question. This ensures that the selected sub-topics can be modified, making the teacher's question-setting operation more flexible and ensuring the efficiency and rationality of the final electronic homework. For example, when the teacher is setting questions, if he thinks that the fourth sub-topic is difficult, or the fourth sub-topic has not been taught yet, he can temporarily not set it and only assign the first three sub-topics. In this way, he can confirm the deletion of the sub-topic by checking the fourth sub-topic. Among them, only one click is needed to check or cancel the selection of the corresponding sub-topic, and there is no need to edit the question document or the answer document in Word (a document format). In addition, similarly, for several sub-topics that have been selected under a question, other sub-topics can be added under the sub-topic as needed until the corresponding question-setting requirements are met.
[0429] In one feasible solution, before the step of constructing the topic data architecture for topic construction, the following steps are also included:
[0430] Using preset decomposition and reconstruction rules to process the preset text content, so as to construct a storage data mapping relationship table for the multi-dimensional analysis question data architecture;
[0431] As shown in FIG24 , a storage data mapping relationship table is constructed by decomposing and reconstructing a certain actual text content; score, source, serial_number, year_start, year_end…are the corresponding data decomposition dimensions, and the overall table formed is the storage data mapping relationship table.
[0432] The specific construction operation is generally implemented through human intervention. Of course, it can also be automatically implemented using technical means that can be automatically implemented according to actual conditions. No specific restrictions are made here.
[0433] Among them, the preset decomposition and reconstruction rules are adapted to the scenario requirements under several different teaching standards;
[0434] The steps to build the topic data architecture for topic construction include:
[0435] Generate the topic data architecture for topic construction based on the storage data mapping relationship table.
[0436] Among them, according to actual needs, specific topics can be selected from the storage data mapping relationship table and arranged in different positions to build the required topic data architecture; different actual needs will result in different topic data architectures. In this solution, the preset text content includes but is not limited to Word document content. By decomposing and reconstructing the Word document content according to the set processing rules, the corresponding data mapping relationship table (or database) is obtained. The data mapping relationship table contains a series of related data for topics and subtopics, including but not limited to: subtopic serial number data storage area / unit, adding topic source labels, etc.; that is, various topics can be stored to completely retain all the information of the original topic, and facilitate standardized topic management and other operations, thereby ensuring the user experience.
[0437] Among them, the advantages of the data storage area / unit are: the sub-topic number is the identity card of the sub-topic, and the identity card can locate the information of which sub-topic (such as good or bad situation, whether the student lost a lot of points), which makes it convenient to quickly locate a certain sub-topic through the identity card and implement teaching management at the sub-topic level; the purpose of adding the source label of the question is to facilitate students and teachers to classify and manage their own various questions, so that they can be easily found when used in the future.
[0438] Specifically, the usage rate of teachers for questions of different attributes can be counted, such as the usage rate of real questions (such as the usage rate of real questions in the past three years) and the usage rate of self-created questions; the teacher's preference for different question banks and questions of different attributes can be obtained through the usage rate, and the system automatically recommends the attributes of questions that teachers with better teaching results tend to use to other teachers; of course, this needs to be executed under the premise of obtaining authorization from excellent teacher users. In addition, some questions can be locked according to the attributes of the questions, such as locking the real questions of last year in the question bank, prohibiting teachers from assigning them in regular homework, and only allowing them to be used in exams. In an implementable solution, the processing method also includes:
[0439] When constructing the question data structure or after obtaining the electronic assignment, obtain the question label editing operation performed by the user;
[0440] In response to a topic tag editing operation, the topic tag data or subtopic tag data of the topic is changed.
[0441] In this solution, when forming an electronic homework, or after a complete electronic homework has been formed, users with operating permissions (students, teachers, schools, education bureaus, etc. all have editing permissions) can click on the edit bar under the corresponding question to edit the label information of the question or any sub-topic under the question, so as to realize the customization function of the label information of all questions; for example, for teachers, after the teacher customizes some labels that are in line with their own usage habits, after a period of time, the teacher can search and find the desired questions according to the customized labels, because the labels are added by the teacher according to his or her own usage habits, and it is easier to remember which labels can be used to retrieve which questions, so that the teacher can view the question bank according to different label categories, and filter out questions that meet specific conditions according to a certain question-setting requirement to assign homework, which ensures the efficiency of the teacher in selecting questions to form electronic homework, and also improves the usage habits and experience on the teacher side.
[0442] For example, each teacher can have their own unique tagging system to suit their categorization habits, making it easier for them to use and search for their work. On the student side, students can label their own completed questions, making it easier to review and revisit them later or try them again, making them easier to use and search for their work. School-wide tags can be standardized and shared by every teacher across the entire school.
[0443] Among them, the purpose of multi-label setting is to facilitate the search of questions. The more labels there are, the more search paths there are, which makes it convenient for teachers to search from any label dimension to ensure the efficiency of question query that meets the requirements, thereby ensuring the overall generation speed of electronic homework.
[0444] Specifically, the system divides the labels of each subject background into three categories: required, high-frequency optional, and low-frequency self-created; of course, they can also be expanded or modified according to actual needs.
[0445] Among them, the purpose of distinguishing three types of tags is to reduce the difficulty of labeling for the first time. You can only fill in the required types. If the required types are not filled in, the question cannot be retrieved. The required types of tags are such as the source of the question labels; the high-frequency optional types are to guide teachers to label this type, but it is not necessary. Teachers are expected to label the questions that meet the requirements with this type of label. The high-frequency optional types of tags are such as knowledge point labels. The types of tags and each specific tag of the low-frequency self-created tag category can be added and edited by the user himself; in order to facilitate the subsequent addition or editing of tags, the low-frequency self-created tag category is set by each teacher according to their teaching habits, and flexibly adapts to each teacher to meet the query habits of the corresponding teachers and meet personal usage needs. In addition, different users can have the same or different low-frequency self-created tags for the same question.
[0446] Of course, for the operating end (students, teachers, schools, education bureaus, etc. all have editing permissions), they can edit labels and have similar operating functions as mentioned above, which will not be repeated here.
[0447] In addition, regarding the determination of question labels: (1) Each teacher can create their own knowledge point labels. These labels may be different from the official labels of the examination board and more suitable for the teacher's own teaching. The teacher can mark the questions using his or her own classification method; (2) After obtaining authorization from excellent teachers, obtain the question label dimensions of excellent teachers and share them with teachers of the same subject; (3) Based on students' previous question-solving behavior, such as which students will make mistakes in which type of questions, sort out a batch of similar questions. These questions may not be consistent in terms of knowledge points, but have similarities in a certain dimension. This batch of questions can be provided to teachers or teaching and research personnel to summarize new question feature dimensions; if this type of question is a question that students are prone to making mistakes in, then a new test ability can be summarized. Teachers can add a new label to these questions and conduct targeted exercises for students.
[0448] Get questions of the same category based on clustering algorithms: automatically find the dimensions that can classify questions, and find the most useful dimensions for classifying questions (the variance on these dimensions is the largest). Based on this standard, you can more effectively form a question set and find similar questions.
[0449] Question feature extraction involves static features such as the subject, the standard grade level for the question (which grade should be taking the question), and text extraction features; dynamic features include the correct and incorrect answers and scores of m students taking the question. Because features have potential connections and certain structure, clusters of the same type are not necessarily distributed in Euclidean space. Furthermore, given the large number of students, a density-based clustering algorithm (such as the HDBSCAN clustering algorithm) is used to achieve good clustering results in non-Euclidean space. This also reduces the impact of outliers and eliminates the need for manual selection of hyperparameters such as the number of clusters.
[0450] Alternatively, the questions can be classified. The first dimension is generally the knowledge point tested in the question, and the second, third, and fourth dimensions are dimensions that were difficult to extract manually in the past. For example, Category A students are good students, and Category B students are above-average students. Based on historical test data, a collection of questions that Category A students can answer correctly but Category B students cannot can be found. These questions may be distributed in different places, but they all reflect the difficulties Category B students encounter in their studies. These questions can be further divided into Category B1 students who can answer correctly but Category B2 students who cannot answer correctly; Category B2 students who cannot answer correctly but Category B2 students who can answer correctly... In this way, students' learning bottlenecks can be accurately found and targeted improvement can be achieved. Among them, this group of questions and the students corresponding to this group of questions are pushed to professional teachers, who will then name the various categories.
[0451] In one feasible solution, the processing method further includes:
[0452] Preset reference answers corresponding to preset sub-questions;
[0453] The preset reference answer includes several sub-scoring points, each of which corresponds to a scoring content and a sub-score, and the sum of the sub-scores is the total score of the sub-question;
[0454] Obtain the student's handwritten answer corresponding to the answer writing track data in the handwritten answer area;
[0455] Identify the actual answer points in students' handwritten answers;
[0456] Compare the actual answer points and the preset sub-score points to get the degree of overlap;
[0457] Based on the overlap, the total score of the sub-questions and the actual sub-score of each actual answer point are obtained;
[0458] In this solution, each sub-question under each question has its corresponding reference answer in advance. The reference answer consists of multiple sub-scoring points, each sub-scoring point corresponds to a sub-score, and the sum of the scores of multiple sub-scoring points is the total score of the sub-question; after a student submits an electronic homework, the overlap between the written content in a certain answer area and the sub-scoring points in the reference answer is obtained. If the overlap is higher than a certain value (such as 95%), it is determined that the corresponding sub-scoring point can obtain the corresponding sub-score, and thus the sum of the sub-scores of the student's actual answer content under each sub-scoring point is calculated (generally based on the teacher's manual correction to determine the answer score, or it can also be Based on the comparison and analysis of machine recognition technology, it is implemented to correct homework, etc., in order to obtain the student's actual score for the sub-question, and to make a detailed division of the statistics of the answer score, so that students can better understand at which sub-scoring points they lost points and at which sub-scoring points they got full marks, and have a clearer understanding of what problems they have in the specific answering process, so that students can understand them in a targeted manner; of course, it is also convenient for teachers to learn about the corresponding score details and provide more targeted guidance; in addition, in situations where there are multiple people marking papers, marking and grading based on the score points can better help individual marking teachers understand the scoring standards and the intention of the questions, so as to enhance fairness.
[0459] Specifically, taking the smallest sub-question as the unit, in addition to entering the question stem and the total score of the sub-question, the answer to each sub-question is entered as scoring points: for example: the total score of the sub-question is 4 points, and it contains three scoring points, namely 1 point, 1 point, and 2 points. What is the correct answer for the first scoring point of 1 point, what is the correct answer for the second scoring point of 1 point, and what is the correct answer for the third scoring point of 2 points.
[0460] In one practicable solution, the plurality of preset sub-scoring points include a first number of substantial sub-scoring points and a second number of insubstantial sub-scoring points;
[0461] Among them, the substantive subscore points and the non-substantive subscore points are arranged according to a preset order;
[0462] Non-essential sub-score points are marked with zero points or no points.
[0463] In this plan, the reference answer to each question includes not only the scoring points, but also the complete answer supplementary content beyond the scoring points. The answer content with the complete answer supplementary content is generally more complete than a student's full-mark answer; among them, the complete answer supplementary content includes but is not limited to listing the corresponding basic solution formulas. The scoring points corresponding to this content are 0 or empty, that is, no score, but students are required to write the answer in a specific place in the answer structure to make the answer more complete, so as to ensure the integrity and rationality of the entire answer content, thereby effectively assisting in standardizing students' answering habits, answering points, etc. to improve overall answering ability.
[0464] Specifically, the 0-scoring point function can help restore the difference between the complete solution process and the scoring points, helping teachers who prepare lessons to understand the problem-solving ideas and question-setting intentions more quickly.
[0465] If only the scoring points of the steps with scores are counted, students may still not understand the specific solution ideas of the problem; although some scoring points are scored as 0, it helps students understand why the problem is solved in this way, thus achieving the effect of assisting and strengthening understanding; students can read the complete problem analysis before class, and if they still have questions, they can solve them in class.
[0466] In addition, the complete analysis process can be used as a model answer, which can help improve students' written expression skills and increase their test scores.
[0467] Specifically, the process of generating, supplementing, and automatically correcting answers to handwritten questions based on AI in this solution is as follows:
[0468] (1) Generate answers based on AI
[0469] The question stem is used as input, and a generative language model (such as GPT) generates an answer array. The answer array consists of several lines, each of which is a step or a scoring point in the answer.
[0470] If there is no score requirement, generate the answer array separately;
[0471] If the user requires a score, the question content and the target total score are input. The answer array is a number of score points, each of which is a combination of content and score. The score of each score point can be 0.5 points, 1 point, 1.5 points, or other values. The score of each score point can also be 0 points. The sum of the scores of each score point is the target total score.
[0472] In actual use, the teacher's workflow may be to set questions first, then mark the homework, but not write the answers; after marking the homework, the teacher converts the correct student answers into rich text format through OCR, and then uses the question stem content and the answers of the students who answered the question correctly as input. The generative language model generates an answer array based on the question stem content and the existing correct answer copies as a reference to improve the accuracy of the generated answers; if the question is relatively difficult and no student in the entire class gets full marks, the answers of the students with the highest score ratio are selected, and the question stem content, the answers of the high-scoring students and the scores are used as input to generate the corresponding answer array.
[0473] (2) Enriching answers based on AI
[0474] The questions teachers usually use include multiple-choice questions that only have correct options but no solution process, and essay questions that only have the final answer or a brief solution process but no complete solution process. Students still cannot understand these questions even if they directly look at the solution. To solve similar situations, the question content and the current simplified version of the answer are used as input, and the generative language model generates the corresponding answer array.
[0475] (3) Automatic correction based on AI
[0476] The question stem, answer array, and student answer image are input into the language generation model, which then issues instructions, including but not limited to: question information: including the question stem, text, formulas (LaTeX format), and images (if any); standard answers: answer content and scoring points; student answers: converted into text and LaTeX-formatted mathematical formulas using OCR technology; and the number of points awarded for asking questions to the large language model.
[0477] Because the output of an untrained large language model may be uncertain, such as responses that do not conform to instructions, the following methods are used to train an automatic grading model that meets the requirements, thereby improving the quality and efficiency of automatic grading:
[0478] Ask GPT the same question multiple times, analyze GPT's answers, and use the most frequently appearing score as the final result;
[0479] Pre-check the quality of student answers, such as whether there are blank spaces, low optical character recognition conversion confidence indicating blurry fonts, etc., and automatically score only those answers that meet the automatic correction conditions;
[0480] Fine-tune the large language model: organize training data, input the question stem and answer content, and output the result of manual correction; use the data set to train the model and adopt LoRA technology to reduce training costs and resource consumption.
[0481] In one feasible solution, the processing method further includes:
[0482] Determine whether the answer viewing conditions are met at the current moment. If so, obtain the student's answer display progress requirements for the sub-questions;
[0483] According to the answer display progress requirements, part of the preset reference answer content corresponding to the sub-questions will be displayed step by step until the complete preset reference answer is displayed.
[0484] Specifically, users input the question and answer, and the generative model is asked to output a detailed explanation. This explanation includes detailed steps and explanations for each step. The explanation is divided into two parts: the first part is a written presentation, such as important derivation formulas and theoretical explanations of relevant knowledge points; the second part is an audio playback, which contains some supplementary explanations, converted from text to audio. The simultaneous presentation of the written step-by-step presentation and the audio playback achieves an automatic explanation effect, providing users with a more convenient and efficient learning experience.
[0485] Among them, the generative model needs to be trained on GPT in advance. The input of the model training is the content and answers of several questions, and the output is the corresponding detailed explanation manuscript. For any question, the system automatically extracts the title and answer on the page and sends them to GPT to automatically obtain the output detailed explanation manuscript.
[0486] In this solution, students can be given the right to view reference answers in certain specific scenarios. These specific scenarios are determined by actual circumstances and can be readjusted based on actual needs. For example, after the teacher has corrected the electronic homework and returned it to the student, or in a scenario where the student completes the homework independently and can compare the answers, the student can click the designated button in the corresponding answer area (that is, the need to view the answer to the question is detected) to view the corresponding reference answer; that is, if the answer viewing conditions are met in other scenarios, the function of clicking to view the reference answer will not be responded to.
[0487] In addition, to ensure that students can think through the questions, the answers are displayed in a step-by-step manner. For example, each click displays the answer to one scoring point until the full answer to the total score corresponding to the question is obtained; or, each click displays 20% of the reference answer content, and after 5 clicks, the complete reference answer to the question is displayed. This not only helps students with reference, but also ensures the quality of supervision and correction, so as to ensure that students have mastered the knowledge points tested in the questions as much as possible.
[0488] Of course, the corresponding reference answers can also be directly displayed in full according to actual needs to ensure the efficiency of students in comparing answers; for example, when estimating the corrected content of the question written by the student, such as by counting the total time the student takes to correct the current question, if the total time exceeds a certain value, it is determined that the student has the need to view the complete reference answer. At this time, when the student clicks the designated button to adjust the corresponding reference answer, the reference answer content will be fully displayed.
[0489] The triggering conditions for enabling the function of viewing answers can be customized or adjusted according to actual needs, making the implementation more flexible and further improving the user experience of the homework answering system. In one feasible solution, the processing method also includes:
[0490] Generate a preset layout template for questions based on the physical assignment style or electronic document assignment style;
[0491] The preset layout template includes several editable areas associated with different subtopics;
[0492] In the existing solutions, the design of the user interface is based on the needs of the computer system program. The corresponding data input boxes are placed on the page according to the data required by the system. This is relatively complicated and difficult for users to operate. It is difficult to meet the user's usage preferences, resulting in users being unwilling to use it at all.
[0493] In this solution, electronic documents such as Word or PDF (a document format) are designed and laid out in a template based on the paper assignment style. This allows teachers to continue using Word editing and printing. The Word-like layout and experience make the editing process less of a change for teachers, ensuring adaptability and ease of use. Furthermore, a preview function is supported, ensuring that the preview interface is as consistent as possible with the actual user interface.
[0494] Treatment methods also include:
[0495] When a first click operation is obtained on an editable area, a corresponding edit box and editing tools are displayed;
[0496] When the editable area is in a default state, or when a second click operation is performed on an area other than the current editable area, the edit box and editing tools of the editable area are hidden.
[0497] Among them, the editing tools include but are not limited to adjusting fonts, inserting pictures and formulas, etc. Only with these functions of the editing tools can the question content be completely stored in the system.
[0498] In this solution, each question corresponds to an editable area. By default, the edit box and editing tools of the editable area are hidden. The user can display the edit box and editing tools on the corresponding area by clicking. When clicking again on any other blank area, the corresponding edit box and editing tools can be hidden, thereby achieving flexible calling of the edit box and editing tools without destroying the overall display effect of an assignment, which is close to the user's original usage habits.
[0499] Specifically, when the questions or test paper are in a non-editing state, the various frame toolbars in each answer area are hidden, making the questions / test paper more like what the teacher originally saw in Word or on paper.
[0500] In one feasible solution, the processing method further includes:
[0501] Get the external input question image;
[0502] Recognize and process the question image and extract the input content;
[0503] The input content includes the question to be input and / or the answer corresponding to the question to be input;
[0504] Generate corresponding sub-topics or questions based on the input content and store them in the preset question bank.
[0505] In this solution, questions can be obtained through external image input. The question image can be an image containing printed text content, an image containing handwritten text content, etc. The existing image recognition technology is used to automatically identify and convert the question and / or answer into an electronic form. In this way, the question can be obtained automatically and efficiently, avoiding the low efficiency of manual entry and the occurrence of errors.
[0506] Specifically, teachers can find the desired questions from books, websites, and other sources, and then directly take a screenshot or save it to obtain the corresponding image. Of course, teachers can also handwrite the questions and answers on paper or a pad, then take a photo or screenshot to obtain the corresponding image. After the system uses automatic optical character recognition technology to translate the questions, the teacher can make fine-tuning adjustments and then store the questions electronically for future reference.
[0507] In one feasible solution, the input content includes text content, image content, and code content corresponding to a calculation formula in an external input image;
[0508] The title image is generated based on handwritten content, printed content, or electronic content;
[0509] The handwritten content refers to the content written on paper materials or electronic handwritten pages.
[0510] In this solution, the input content can include different types, such as codes corresponding to text, pictures, and formulas. For example, for a physics handwritten question, the question is composed of text, pictures, and formulas. While the text and pictures can be recognized and extracted, the formulas can also be extracted with high quality. For example, the code language corresponding to y=ax is preset to "0101". When the formula y=ax exists in the image, it is automatically recognized and directly converted to "0101" for input. In the set scenario, such as when the teacher selects the question from the question bank or when the student does homework, "0101" is automatically converted to y=ax for display, thereby realizing automatic recognition and extraction of all questions, and the process of entering questions into the question bank is fully automated, ensuring the efficiency and quality of electronic homework acquisition.
[0511] Among them, the entered contents can generally be adaptively edited to meet the requirements of question entry; it should be noted that if the formula is entered based on a screenshot, there is no way to edit it later; for formulas converted into electronic code format, they can continue to be edited.
[0512] In addition, you can write the questions and answers by hand on paper or a pad, and then take a photo or screenshot to get the corresponding image. For example, when the teacher's handwritten problem-solving process (mathematical and physical formulas + diagrams) needs to be entered into the system, there is no need to retype the problem-solving process. The system automatically recognizes and processes the image and converts it into electronic storage, ensuring the processing efficiency and convenience of the entire entry process.
[0513] In one feasible solution, before the step of selecting a number of questions from a preset question bank based on the preset learning objectives, the step further includes:
[0514] Construct several topics;
[0515] Receive the user's check result on whether there are preset errors in the question;
[0516] When the check result indicates that there is a preset error, a prompt message is generated to prompt the user to correct the current question until a question without the preset error is obtained;
[0517] When the check result indicates that there is no preset error, the corresponding question is stored in the preset question library.
[0518] Pre-set errors include, but are not limited to, clerical errors and academic errors. Specifically, clerical errors include typographical errors and typos, while academic errors include omitted conditions and factual errors.
[0519] Specifically: for different preset errors, targeted detection technology can be used; for copy errors, typo checking technology can be used;
[0520] For academic errors such as x+x=3x, or misplacement of questions and answers, for example, the answer to question 1 is entered in the answer position of question 2, the answer to question 2 is entered in the answer position of question 3, and so on. Similar problems are detected based on the following methods:
[0521] For each question and each sub-question, a data grid indicates the question's inspection status: not automatically inspected, passed automatic inspection, passed manual inspection, or incorrect. Initially, every question added to the system is in the "not automatically inspected" state. After a period of time, these newly added questions that have not been automatically inspected are input into the multi-layer perceptron neural network model for inspection. The neural network outputs a question credibility variable between 0 and 1. The closer the question credibility variable is to 1, the greater the probability that the question is incorrect. Questions with question credibility variables exceeding the set threshold are submitted for manual inspection to reduce the workload of manual inspection. After manual inspection, these questions are marked as either "passed manual inspection" or "incorrect." The algorithm also adjusts the weights of the parameters in the neural network based on the results of the manual inspection to improve the accuracy and efficiency of the system's automatic inspection.
[0522] In this solution, in order to ensure the final quality of electronic homework, it is necessary to ensure that there are no problems with the questions in the most basic question bank; increase the quality check of the questions to determine whether there are errors in each question, and remind teachers to make targeted corrections to make them correct when errors are verified; for questions that are directly determined to have no errors, they are directly stored in the question bank for subsequent retrieval.
[0523] Specifically, after the front-line question recorders (full-time staff, teachers, etc.) record the questions, errors are inevitable; if the entered questions are past years' real questions in the question bank, if there are no quality inspection questions, and there are errors in the questions, when other users see them, it will affect other users' usage experience and trust in the answering system in this solution; therefore, only by configuring the quality inspection function and having dedicated inspectors to carefully verify, and click to pass after confirming that there are no errors, and then put them into the final question bank, can it be guaranteed that the questions called by subsequent users are correct and there will be no wrong questions.
[0524] In one feasible solution, the processing method further includes:
[0525] Pre-build several electronic question bank folders;
[0526] As shown in FIG25 , the same electronic question bank folder corresponds to a number of questions corresponding to the same subject in the same grade;
[0527] During the teaching process, it's necessary to standardize the construction and management of question banks to ensure efficient question selection and improve the efficiency of obtaining electronic homework. For example, it's necessary to use certain labels to distinguish and store questions used in Grade 12 Math from those used in Grade 13 Math, and vice versa. In some special cases, even if the subject and grade level of the course content are the same, question banks still need to be differentiated, such as the question bank for the Grade 12 Liberal Arts Math paper and the question bank for the Grade 12 Science Math paper in the Gaokao.
[0528] Different classes in the same grade, teaching the same subject, and covering the same content typically prepare lessons together, sharing the same set of materials and homework. For each subject, a pre-built electronic question bank folder is typically used to store all available questions for that subject, making it easier to search or select from it to create electronic homework assignments. New questions, such as teacher-created questions, can also be added to the corresponding electronic question bank folder as needed for updating.
[0529] When preparing lessons, teachers add appropriate questions to the corresponding electronic question bank folder. When adding, you can add labels to each question according to your own habits, such as chapter unit labels, knowledge point labels, difficulty labels, etc., so that they can be called up in time when searching for questions in multiple dimensions. You can also add dimensions to the question labels yourself, such as adding a dimension to whether the question is locked. Questions marked as locked will be used in exams. Teachers use this label to determine whether the question is used for daily homework practice or for exams, thereby greatly improving the efficiency of electronic homework formation and improving the teacher's experience of using the homework management system.
[0530] The homework management system can also conduct further statistical analysis on these label data corresponding to each question. For example, according to the chapter unit label, it can count how many questions there are in each chapter unit in the current electronic question bank folder, and whether any chapter unit needs to add more questions; according to the difficulty label, it can count the proportion of questions of different difficulty levels to determine whether the questions currently added to the electronic question bank folder are appropriate, so as to facilitate timely adjustment. The statistical analysis process is fully automatic and does not require manual calculation, ensuring processing efficiency, accuracy, and convenience.
[0531] For any subject group in a school, the teaching management staff can flexibly set the rules for establishing different electronic question bank folders and which classes correspond to the same folder. As long as the teaching content of several classes is the same, the same electronic question bank folder can be reused and can be called up and reused in the next academic year or even in subsequent academic years.
[0532] The process of building and improving the question bank can be completed collaboratively by several users; it can also be centrally built by the school teaching and research group before the start of the semester, and then handed over to specific teachers. New questions can be continuously added during the semester, and less suitable questions can be eliminated to gradually improve the quality of questions. The question bank built in the first year can also be continuously updated according to actual conditions and teaching results in the subsequent teaching process each year.
[0533] The information technology platform provides a feasible and efficient foundation for building a question bank. Using traditional document-based work models, it's difficult to achieve collaborative updates among multiple users. Managing the question bank across multiple academic years can also lead to document confusion or loss, preventing valuable firsthand teaching feedback and teacher experience from being captured and consolidated. The information technology platform provides the foundation for gradually improving question quality, enhancing the quality of students' regular practice assignments. This allows students to achieve their learning goals with fewer questions, reducing their workload and significantly optimizing existing teaching practices.
[0534] Get the teaching related parameters of each subject in each class of the same grade;
[0535] Among them, teaching-related parameters include different teaching progress corresponding to different lesson preparation contents.
[0536] Generate several initial electronic homework assignments for each class based on the teaching-related parameters and the questions in the electronic question bank folder of the corresponding subject;
[0537] As shown in FIG26 , several initial electronic homeworks for each class are stored in the same homework folder, all of which are from the same electronic question bank folder; different initial electronic homeworks correspond to different preset teaching schedules;
[0538] Based on several initial electronic assignments, the target electronic assignments to be assigned each time are determined and published to the student end.
[0539] Specifically, a class will have multiple electronic assignments assigned during a semester, and these assignments need to be organized and collated. For example, a teacher teaches four classes simultaneously, assigning two assignments per class per week. This translates to 160 different electronic assignments over the course of a semester. The teacher needs a more convenient and efficient way to manage and store these assignment documents.
[0540] When a user first enters the electronic library folder, he or she can choose to add questions manually or import questions automatically. The automatic import module performs the import operation in the following way:
[0541] Method (1): The user inputs the total number of questions that the question bank wants to have and the actual exam questions from which year to which year they want to use; the knowledge point outline corresponding to the subject background is read, and n question bank folders are created according to the first-level directory of the knowledge point outline; if the number of the first-level directory is less than 5, the subject background knowledge point outline is not designed in such a detailed manner, then the question bank folders are created according to the second-level directory and the third-level directory in turn;
[0542] The system scans all real exam questions within the subject context and time range in the system, obtains the knowledge points corresponding to each question, and determines the ratio of the number of questions for each knowledge point to the total number of questions; based on this ratio, the number of questions in each knowledge point folder is determined, and the questions corresponding to each knowledge point are added to the corresponding folder; the selection criteria: random selection can be performed, or questions with the most overall selections by teachers can be added first based on the usage history of other users in the system; after the initial question bank is generated, the initial question bank can be displayed to the user, and the user can be asked whether the user needs to modify the expected total number of questions or whether he wants to replace all questions;
[0543] Method (2): Learn from other teachers' question banks. Specifically, the system asks the teacher user whether he is willing to share his question bank resources with other teachers of the same subject. Among the teachers who agree to share, the system will choose whether to adopt the teacher's question bank based on the teacher's teaching ability assessment; by learning the folder classification methods of many excellent teachers, establish question bank folders; learn the question selection of many excellent teachers, and combine the questions they selected into various knowledge point folders.
[0544] After that, enter the refinement step. There are multiple modes for the refinement step. The following two modes are taken as examples:
[0545] Mode 1: The user selects which folders need adjustment and which do. A new batch of questions is then selected for the folders that need adjustment. Mode 2: The user deletes questions they dislike in each folder. The system then automatically adds the same number of new questions, repeating this process until the user is satisfied. During this process, the "goodness" of each question is updated based on feedback from multiple users, optimizing the quality of future first-time recommendations.
[0546] In addition, the implementation process of the question bank evaluation and suggestion module in the system is as follows:
[0547] The user has created his own electronic question bank folder. By using an artificial intelligence model to analyze the following data, the following suggestions can be given. Each suggestion is accompanied by an associated question. Teachers can directly select questions from the suggested associated questions and add them to the lesson preparation folder, or they can select the questions they want themselves.
[0548] Suggestion 1: Compare the outline of the background knowledge points of the exam, create a folder with the teacher, and give suggestions: "X chapter knowledge points are not covered, should this lesson preparation folder be added?"
[0549] Suggestion 2: Automatically calculate the percentage of the question bank occupied by each knowledge point, compare it with the actual number of questions given by the teacher, and give suggestions: "Should we add more questions to this chapter?"
[0550] Suggestion 3: Based on the teaching data from the previous academic year or several academic years, identify the knowledge points that were not taught effectively in the past and provide suggestions: "Should more questions be added to this chapter?"
[0551] Recommendation 4: Based on historical exam data, provide suggestions: "This chapter's questions are too easy, should we add more difficult ones?" "This chapter's questions are too repetitive, should we delete them?" Based on historical exam data, if the average difficulty of this knowledge point in real history exams is relatively high, remind the teacher, "The historical questions on knowledge point XXX are quite difficult, should we increase the amount of practice for this chapter?"
[0552] The system also has a difficulty variable setting function, the specific implementation process is as follows:
[0553] Each question has a difficulty variable x, which indicates its difficulty. The difficulty variable is dynamically calculated during the problem-solving process. The difficulty variable is a value between 0 and 1. Initially, the difficulty variable for all questions is set to 0.5. The theoretical grade level corresponding to each question is also determined, such as fourth grade or senior year. Each time a student completes a question, the difficulty variable is updated based on whether the student completed the question.
[0554] The system also has a dynamic update function, the specific implementation process is as follows:
[0555] Based on the student's relative ranking among all students in the grade, a student ability variable y between 0 and 1 is calculated; where an ability variable of 1 represents that the student is ahead of all students in the grade, and an ability variable of 0 represents that the student lags behind all students in the grade.
[0556] If students try to solve problems in different grades, if they try to solve problems in a higher grade, such as a third grader trying to solve problems in a fourth grade, then the student's ability variable should be adjusted to y = y / 2; if they try to solve problems in a lower grade, such as a fourth grader trying to solve problems in a third grade, then the student's ability variable should be adjusted to y = (1+y) / 2;
[0557] If the student ability variable is higher than the question difficulty variable, the student will make mistakes in the question, and the question difficulty variable will be increased, x=x+(yx)*k, where k is a fixed coefficient, such as 0.001.
[0558] If the student ability variable is lower than the question difficulty variable, the student answers the question correctly, which reduces the question difficulty variable, x = x-(xy)*k, where k is a fixed coefficient, such as 0.001.
[0559] In this solution, the generated lesson preparation folders are analyzed automatically, reducing the time teachers spend on lesson preparation, thereby improving the quality of homework and improving students' learning efficiency.
[0560] Teachers can create several new electronic assignments for each class, pre-setting the name and expected release date of each electronic assignment. At this time, each electronic assignment is blank. Then, they can add questions from the electronic question bank folder to each specific electronic assignment in different ways. During the editing process, assignments that have been edited to different degrees can be temporarily saved as drafts, and gradually improved in several stages. Finally, a confirmed version of the assignment is obtained as the initial electronic assignment. In this way, the standardized archiving management of all electronic assignments is guaranteed. After the end of a semester, teachers can come back and review several assignments assigned to a class in the same semester. The system will automatically complete the archiving work, which greatly simplifies the teacher's handling of electronic assignment management and reduces the time invested by teachers. As time goes by, the advantages of this standardized archiving management solution for electronic assignments become more prominent.
[0561] In one feasible solution, the processing method further includes:
[0562] Preset the first question search interface in the teacher's terminal;
[0563] The first search interface includes a plurality of first search boxes, and the search contents in different first search boxes correspond to search parameters of different dimensions; or, the search contents in the first search box include a preset portion of the question stem content or a preset portion of the answer content;
[0564] The steps of generating a plurality of initial electronic assignments corresponding to each class according to the teaching association parameters and the questions in the electronic question bank folder of the corresponding subject include:
[0565] According to any preset teaching progress, determining a first actual search content in the question search interface to obtain a plurality of matching first questions from an electronic question bank folder of a corresponding subject;
[0566] A first set number of target topics are selected based on the first topics to form an initial electronic homework for the corresponding class, until a number of initial electronic homeworks corresponding to different preset teaching progresses are obtained.
[0567] Table 1
[0568] In this solution, as shown in Table 1 above, each subject's electronic question bank folder generally contains a large number of questions. Teachers can find suitable questions based on students' learning progress and learning goals. For example, the difficulty dimension of a question can be labeled as 1 star, 2 stars, 3 stars, 4 stars, or 5 stars; the knowledge point dimension of a question can be labeled as addition, subtraction, multiplication, and division; and the attribute labels of a question can be single-choice or multiple-choice. Question A has the labels: 2 stars, addition, and question B has the labels: 3 stars, addition, and multiplication. The question bank also contains other questions. Users can first search for "knowledge point addition" in the search box of the question search interface, view the search results, and add several questions. Then, they can search for "difficulty 1 star + knowledge point subtraction + multiple-choice" in the search box of the question search interface, view the search results, and add several questions. This search method is used when providing suggestions and suggesting related questions, such as searching by knowledge point or searching for knowledge point + difficulty.
[0569] In addition, if the teacher remembers that there is part of the stem of a question in the question bank folder, he or she can directly search for questions based on the corresponding content; for example, he or she can enter "Xiao Ming has 5 apples" and compare and match the text content to search for more suitable questions.
[0570] In this solution, teachers can search for questions directly in the question bank based on the label information of different dimensions of the questions, or based on the content contained in the question stem or the reference answer of the question, so as to achieve the effect of efficiently and quickly searching for related questions.
[0571] In one feasible solution, before the step of determining the target electronic assignment to be assigned each time based on the plurality of initial electronic assignments and publishing the target electronic assignment to the student terminal, the method further includes:
[0572] Selecting the first class in the same grade in previous years whose correlation with the teaching correlation parameter of the current class is greater than a first set threshold;
[0573] Retrieve the historical related homework of the same teaching progress published in the first class;
[0574] Use the History Connections assignment as the initial electronic assignment for the current class; or
[0575] Selecting a second set number of topics from the historical related homework as target topics, and selecting a third set number of topics from the first topics obtained through the search box as target topics, so as to obtain initial electronic homework for the corresponding class;
[0576] or,
[0577] Selecting a second class in the same grade in the current year whose correlation with the teaching correlation parameter of the current class is greater than a second set threshold;
[0578] Retrieve the first electronic homework of the same teaching progress from the homework folder of the second class;
[0579] Use the first electronic assignment as the initial electronic assignment for the current class;
[0580] Selecting a fourth set number of questions from a first electronic homework and selecting a fifth set number of questions from the first questions obtained through the search box as target questions to obtain an initial electronic homework for the corresponding class;
[0581] A sixth set number of questions are selected from the plurality of first electronic assignments as target questions to obtain an initial electronic assignment for the corresponding class.
[0582] In this solution, teachers can obtain the electronic homework assigned for the same subject in the previous class (or even earlier) at the same period (i.e., the same teaching progress) based on the actual teaching progress of the current class, and then use the electronic homework as the electronic homework for the current class. They can also add or delete questions based on this basis to obtain electronic homework suitable for the current class.
[0583] The process of automatically obtaining a teacher's teaching progress is as follows: For example, in the 2022-2023 academic year, a teacher assigns five assignments to a class: Class A Homework 1, Class A Homework 2, Class A Homework 3, Class A Homework 4, and Class A Homework 5. For a class with the same content in the 2023-2024 academic year, when the teacher first enters the class to assign homework, the system automatically prompts: "Do you want to assign Class A Homework 1?" If the teacher has already assigned homework in the class, the system automatically calculates that the content of Class B Homework 1 overlaps 80% with Class A Homework 1, and that the content of Class B Homework 2 overlaps 80% with Class A Homework 2. If the teacher enters the system later, the system automatically prompts: "Do you want to assign Class A Homework 3?" The teacher then makes fine-tuning based on this to ensure efficient homework assignments.
[0584] In addition, for example, if a teacher teaches two classes at the same time, and Class 1 progresses slightly faster than Class 2, when the electronic homework for Class 2 is issued, the teacher can select an electronic homework from the electronic homework assigned to Class 1, and then add or delete questions on this basis to obtain the electronic homework suitable for the current Class 2.
[0585] Specifically, for the same subject, the relevance of each other class to the current class is calculated, and the class lists corresponding to each class are displayed in descending order according to the size of the relevance. This makes it easier for teachers to select the class with the greatest relevance to the current class and the electronic homework assigned during the same period from the class list.
[0586] As shown in Figure 27, for example, the history class list of mathematics for Class 4 (1) in the 2023 academic year can be: mathematics for Class 4 (2) in 2023, mathematics for Class 4 (3) in 2023, mathematics for Class 4 (1) in 2022, mathematics for Class 4 (2) in 2022, mathematics for Class 4 (3) in 2022, mathematics for Class 5 (1) in 2023, mathematics for Class 5 (2) in 2023, mathematics for Class 5 (3) in 2023, mathematics for Class 3 (1) in 2023, mathematics for Class 3 (2) in 2023, mathematics for Class 3 (3) in 2023; the class with the greatest relevance is matched: mathematics for Class 4 (2) in 2022, and then the 9.1 electronic homework assigned in the same period is selected and Question 1 and Question 2 are selected as part of the electronic homework for mathematics for Class 4 (1) in the 2023 academic year.
[0587] In addition, cross-grade electronic homework can be used for reference. For example, if the fourth-grade math class needs to use a question from the fifth-grade math class, the teacher can also retrieve it through the historical class list to achieve the purpose of calling the corresponding question.
[0588] In this solution, teachers can call up available homework for reference in a timely manner without having to create new questions for each electronic homework, which greatly reduces the teacher's workload. At the same time, by drawing on previous homework, the quality of electronic homework can be improved to a certain extent.
[0589] In one feasible solution, the search interface includes a question overview area and a question detail preview area;
[0590] After the step of obtaining a plurality of matching first questions from the electronic question bank folder of the corresponding subject, the method further includes:
[0591] Control the display of a list of questions corresponding to the first question in the question overview area in order;
[0592] The step of selecting a first set number of target topics based on the plurality of first topics to form an initial electronic homework for the corresponding class includes:
[0593] When one or more first topics are selected, the control displays the topic content of the selected first topics in the topic detail preview area, and determines whether to use the selected first topics as target topics according to the actual selection operation, until a first set number of target topics are obtained, so as to form an initial electronic homework for the corresponding class;
[0594] In this solution, based on the user's search requirements, several questions are searched from the question bank folder of the corresponding subject, and then these questions are sorted in sequence and displayed in a list. At this time, only the stem content of the beginning of each question is displayed in the list; you can delete the corresponding question by clicking the delete plug-in at a question in the question overview area, or right-click and select delete to delete the corresponding question. When the user checks one or more questions, the checked questions are expanded and displayed in the question details preview area in sequence to display the content of each question in detail; at this time, these questions can be displayed in the question details preview area according to the order in which they are checked, or they can be displayed in the question details preview area according to the order in which they are checked. This can be predetermined or adjusted according to actual needs. You can also click a specific plug-in at any question location in the question details preview area to control the display, such as controlling the display of answers for all questions, hiding answers for all questions, expanding all questions for viewing, or folding all questions to display only two lines of preview, etc.
[0595] For example, if 10 questions that meet the requirements are found in the question bank, these 10 questions will be displayed in a list in the question overview area first, and the question number of each question will be automatically calculated and displayed on the left to avoid the tedious and error-prone manual editing of question numbers. No questions will be added to the currently edited assignment first. When the user checks questions 3, 5, and 7 among the 10 questions, or selects all questions with one click, the selected questions will be added to the question details preview area for expanded display, and the teacher can edit the assignment questions in the question details preview area.
[0596] In this solution, through the clever and reasonable setting of the functional areas in the search interface, teachers can flexibly search, view, select, edit and perform other operations, thereby ensuring the efficiency of electronic homework formation while improving the user's operating experience.
[0597] In one feasible solution, the processing method further includes:
[0598] Based on the first question in the question details preview area, the number of questions and the corresponding total scores are summarized and displayed to determine whether the assignment formation requirements are met;
[0599] Specifically, for example, the full score of the corresponding question is marked on the far right of the last row of each question; the system counts the current assignment, the total number of questions, and the total score, and displays it for the user's reference; among them, the user can adjust the score of each question in the electronic assignment. For example, according to the originally entered scoring points, the total score of a question is 4 points, but the user can adjust the score of this question in this paper to 5 points.
[0600] In this solution, before generating an available electronic homework, the number of selected questions and the corresponding total score of the questions can be automatically calculated, so that the teacher can determine whether to continue selecting questions for the current electronic homework, whether to replace certain questions, or modify the score of a certain question, until the number of questions in the electronic homework is within the teacher's expected processing range and the total score of the questions is 100 points. The number of questions and the corresponding total score of the questions are displayed in a specific display area of the search interface for the user to view. The entire process is completed automatically, ensuring the convenience and efficiency of obtaining electronic homework. In one feasible solution, at least one of the following operations is performed on different first questions in the question details preview area:
[0601] Adjust the order of several first questions, edit the related content of any first question, delete one or more first questions, and mark the question numbers of several first questions in sequence;
[0602] In this solution, the order of different questions in the question details preview area can be adjusted by dragging and dropping, and the label information of any question can be modified or added or subtracted. When checking the checked questions and confirming that they do not meet the requirements, one or more questions can be deleted one by one or selected and deleted at once. Several questions in the question details preview area can also be automatically sorted, numbered and displayed in sequence from top to bottom, etc. Any question in the question details preview area can be flexibly set to meet the various personalized homework formation needs of different teachers.
[0603] In an implementable solution, based on the selected display mode switching instruction, the title contents of the plurality of first titles in the title detail preview area are displayed using a corresponding display mode.
[0604] In this solution, multiple questions in the question details preview area can display only the question stem content, or can display the question stem content and reference answers of each question at the same time. One display mode can be used as the default display mode. If the user needs to switch the display mode, he can switch the display by clicking a specific plug-in in the search interface. The switching display can be a one-click switching display for all questions in the question details preview area, or it can be a display switching for any question in the question details preview area, thereby meeting the display control needs of different teachers.
[0605] In one feasible solution, the step of selecting a first set number of target topics based on a plurality of first topics to form an initial electronic homework for the corresponding class includes:
[0606] A first set number of target topics are selected based on the first topics, and initial electronic assignments are generated for the corresponding class using a physical assignment format or an electronic document assignment format;
[0607] In this solution, electronic assignments are generated in the form of physical assignments or electronic document assignments, following the assignment templates that students and teachers are accustomed to. They are basically close to paper test papers, ensuring the usage habits of students answering questions and teachers correcting assignments, and ensuring the user experience.
[0608] In one feasible solution, the processing method further includes:
[0609] Get the actual user operation on the first search interface:
[0610] Based on the actual operation of the user, obtaining multi-dimensional identification information of the corresponding first question;
[0611] Determine whether to use the current first topic as the target topic according to the identification information;
[0612] In this solution, by obtaining the user's click operation in the question details preview area, the relevant information of any question can be expanded to view, such as which test papers the corresponding question has been used in before, the answer sheets of students who have previously done this question, the historical error rate of this question, etc., to help teachers select higher-quality practice questions, while reducing the burden on students and achieving more efficient teaching goals.
[0613] After the steps of generating the initial electronic assignments for the corresponding class, it also includes:
[0614] According to the identification information of each target topic, the identification parameters of the initial electronic assignment are constructed.
[0615] The identification parameters of the electronic assignment include the assignment name, assignment mode, etc. The assignment mode can have multiple settings, including but not limited to assignment mode and examination mode.
[0616] When the identification parameter includes the homework mode and the homework mode is the examination mode, the sub-score of each target question and the total score of the entire initial electronic homework and / or the corresponding fixed answering time are identified;
[0617] Among them, the fixed answering time is used to provide a countdown reminder in the exam mode, and trigger the submission of the corresponding electronic assignment when the countdown reaches zero;
[0618] When the identification parameter includes the homework mode and the homework mode is the after-school homework mode, the initial electronic homework is not marked with a score and / or the corresponding recommended answering time is not determined.
[0619] In this solution, as shown in Figure 28, in Assignment Mode, users can set the release time for assignments. Students cannot access the test paper before the release time. Users can also set a deadline for assignments, and students are required to submit their assignments before the deadline. The system will remind students to complete and submit their assignments on time based on the deadline. It will also use the deadline to count which students submitted their papers on time and which students submitted their papers late. Assignments in Assignment Mode have no default time limit, but users can choose to add a time limit. Students must complete the assignment within the specified time, and the assignment will be automatically submitted when the deadline is reached.
[0620] In homework mode, users can choose whether to grade their homework. If not set, the score for each question and the total score will not be displayed. When grading, teachers only need to mark the answers and add comments without assigning any grades. If graded is set, the score for each question and the total score will be displayed. Teachers will need to assign a score for each question and each sub-question when grading.
[0621] The system will pre-calculate an initial duration based on the total score for the user's reference. The user can adjust the duration of the corresponding homework / exam based on this. The calculation of the initial duration is determined by the teaching content. For example, for the college entrance examination, the ratio of 150 points to 120 minutes is used for conversion.
[0622] As shown in Figure 29, in exam mode, the test paper has a countdown; the user must set the total exam duration, and students must complete the test within the specified time. Exams in exam mode have open time periods. Only during these open time periods, or exam time periods, can students view the test paper and answer questions. At other times, students cannot view the test paper or answer questions.
[0623] In one feasible solution, the processing method further includes:
[0624] When the first topic is determined as the target topic and the identification information of the first topic indicates that the first topic has been selected and formed into an electronic homework and issued to the current class within a set time period from the current moment, a prompt message is generated.
[0625] In this solution, for this electronic homework, the selected questions are automatically marked. If a question is assigned repeatedly, a reminder message is automatically generated to require manual judgment again. The user can decide whether the same question needs to be reused. Otherwise, the question will be manually deleted, thus avoiding the situation where the same question often appears in the same homework in traditional homework assignment scenarios, and ensuring the quality and accuracy of electronic homework questions.
[0626] In one feasible solution, the processing method further includes:
[0627] Save the newly added questions into the electronic question bank folder of the corresponding subject, or into the temporary electronic question bank folder.
[0628] In this solution, when no suitable questions are found among the questions searched in the question bank, the user can add questions he or she created to this assignment; based on the user's (the teacher himself or herself, other teachers, subject leaders, etc.) judgment of the question, he or she can choose whether to add the question to the question bank folder corresponding to the teaching content, or just save the question in the user's personal temporary question bank folder so that the user can reuse it later, ensuring the timeliness of the collection and storage of new questions, effectively expanding the questions in the original question bank folder, and making it easier for subsequent users to better select questions that meet their needs.
[0629] In one feasible solution, the step of determining the target electronic job to be arranged each time based on the plurality of initial electronic jobs includes:
[0630] Get the actual teaching progress of the current subject in each class;
[0631] Comparing the actual teaching progress with the preset teaching progress of the initial electronic homework of the corresponding subject to obtain a comparison result;
[0632] Based on the comparison results, the designated topics or subtopics under the designated topics in the initial electronic assignments are adjusted to form target electronic assignments that match the actual teaching progress of each class.
[0633] In this plan, when preparing lessons, the teacher originally selected 10 questions to form the initial electronic homework. After class, the teacher fine-tuned the questions in the initial electronic homework based on the actual class progress to obtain the target electronic homework that can be published; among them, questions can be deleted as a unit of major questions, or as a unit of sub-topics; or questions can be added as a unit of major questions, or as a unit of sub-topics.
[0634] In one feasible solution, the steps of adjusting the designated topic or subtopics of the designated topic in the initial electronic assignment based on the comparison results to form a target electronic assignment that matches the actual teaching progress of each class include:
[0635] If the comparison result indicates that the actual teaching progress is slower than the preset teaching progress included in the lesson plan, then the designated questions or sub-questions under the designated questions corresponding to the unfinished knowledge points in the preset teaching progress will be deleted from the initial electronic homework to form a target electronic homework that matches the actual teaching progress of each class; or, after deleting the questions from the initial electronic homework, according to the external demand for adding questions, the corresponding questions will be added to the initial electronic homework to form a target electronic homework that matches the actual teaching progress of each class;
[0636] If the comparison result indicates that the actual teaching progress is faster than the preset teaching progress included in the lesson plan, then the designated questions or sub-questions under the designated questions corresponding to the knowledge points that exceed the preset teaching progress will be added to the initial electronic homework to form a target electronic homework that matches the actual teaching progress of each class;
[0637] Among them, the actual teaching progress corresponds to a series of knowledge points that need to be newly mastered, and the preset teaching progress also corresponds to a series of knowledge points that need to be newly mastered. If the actual teaching progress and the preset teaching progress correspond to the same range of knowledge points that need to be newly mastered, then the two are determined to be consistent; otherwise, the actual teaching progress corresponds to more knowledge points that need to be newly mastered than the preset teaching progress, then the actual teaching progress is determined to be faster than the preset teaching progress; otherwise, the actual teaching progress is determined to be slower than the preset teaching progress.
[0638] In this solution, there is a high probability of inconsistency between the preset teaching progress corresponding to the lesson preparation and the actual teaching progress each time, which may be fast or slow. At this time, the questions in the initial electronic homework need to be adjusted; for example, question 10 has 4 sub-questions, and the actual teaching progress is a little slower than the preset teaching progress, and the knowledge point of the 4th sub-question is not taught. At this time, the 4th sub-question of question 10 needs to be deleted; the deletion operation is also reversible, and the deleted sub-questions can be restored, and then the total number of sub-questions of this question can be checked, and the selection of deletion, addition or modification can be made again to ensure the flexibility and rationality of the electronic homework setting.
[0639] In one feasible solution, the processing method further includes:
[0640] Preview or edit the target electronic assignment before publishing it on the teacher's end;
[0641] In this solution, before publishing the electronic homework, the teacher can preview the status of the electronic homework from the student's perspective and check whether there are any errors in the test paper. If there are errors and other places that need to be modified, the teacher can re-enter the editing state and can edit multiple times to ensure the convenience of the teacher in handling electronic homework.
[0642] The steps to publish the target electronic assignment to the student terminal include:
[0643] Before the teacher releases the target electronic homework and sets the release time in advance, the system time is monitored, and when the system time reaches the release time, the target electronic homework is released to the student terminals of all students in the current class.
[0644] Specifically, after the teacher uploads the class schedule in PDF, Excel, Word and other formats, a language analysis model (trained based on chat GPT) is used to analyze the schedule content to obtain the teacher's class time. For example, the model automatically outputs the teacher's class time as Monday, Wednesday and Friday. Based on the schedule time, the current assigned homework is calculated and the deadline for the next homework is automatically generated.
[0645] Based on the chat GPT model, we learn the time pattern of teachers assigning homework, as shown in Figure 30. For example, for the first four homework assignments of this class, the teacher set the deadlines at 8:30 on Monday, 8:30 on Wednesday, and 8:30 on Friday. The deadline is 8:30 on Monday. When the second Monday appears, with a cycle of 3, the deadline of the subsequent homework assignments will be automatically set to 8:30 on Wednesday and 8:30 on Friday through model training and learning.
[0646] In this solution, users can choose whether to publish immediately or with a delay according to actual needs. For example, for an exam that starts at 9 am on Friday, the teacher can publish the test paper to the student end at that time. Alternatively, if the teacher has completed the test paper on Thursday morning, the test paper can be published in advance and scheduled for publication at 9 am tomorrow. The teacher does not need to make a temporary operation the next morning. For test papers set up in this way, students will only receive the test papers in their accounts when the appointment time arrives, and will not see the test papers before that time, so as to ensure the flexibility and convenience of the teacher's end in publishing electronic homework. Of course, in order to facilitate the student end to receive the electronic homework in the exam mode in time, a reminder can be sent to the student end five minutes before the exam, and a notification message informing that the electronic homework in the exam mode can be opened can be sent simultaneously when the homework is published, thereby ensuring the smoothness of online teachers.
[0647] In one feasible solution, the processing method further includes:
[0648] Obtain the actual academic information of any subject of the target student;
[0649] Determine, based on actual learning information, a second actual search parameter input into the first search box in the question search interface, and obtain a plurality of matching second questions from the electronic question bank folder;
[0650] A seventh set number of questions are selected from the plurality of second questions as target questions to form target electronic homework that matches the actual homework needs of the target students for the corresponding subject.
[0651] When a student completes an initial assignment, for example, questions 1, 4, and 5 are correct, but questions 2 and 3 are wrong, the teacher can assign further consolidation exercises based on the wrong questions, or the system can automatically push consolidation exercises to the student.
[0652] The method of pushing questions is as follows:
[0653] Solution (1): Based on manual implementation, the teacher selects question 1 in advance, and then has two consolidation exercises, 1a and 1b; question 2, has two consolidation exercises, 2a and 2b; and so on. Based on the mistakes made by each student in the first homework, consolidation exercises are pushed;
[0654] Solution (2): Automatic push method. Specifically, based on each wrong question of the student, find questions with the same label as the original question but that the student has not done; or find questions with a certain degree of similarity to the original question and automatically push additional exercises; based on the student's performance in doing the additional exercises, decide whether to push again until the student masters the knowledge points of the corresponding question.
[0655] In this solution, the need to search for questions is determined based on the actual learning situation of any student or multiple students in each subject, such as which knowledge points are mastered well and which are mastered averagely or poorly. The corresponding questions are searched from the question bank and electronic homework is generated to suit the students' weaker knowledge points in the subject, so as to achieve high-quality homework assignments and reduce the burden on students and teachers.
[0656] Specifically, electronic homework can be published as a class unit, or as other units (such as some students, a certain student, etc.). When you choose to publish as a class unit, all students who are set to belong to the corresponding class will receive the test paper collectively. If it is published as a single student, or a group of students with similar learning conditions, the user selects the questions by searching the question bank, selecting questions from historical homework, or making up their own questions, and then publishes them to designated users. For example, if a student has obvious problems mastering a single knowledge point, or is absent from class due to illness or personal leave and needs additional learning assistance, the teacher can form certain questions into electronic homework and assign them to the student separately.
[0657] In one feasible solution, the processing method further includes:
[0658] Constructing a second question search interface in the student terminal or the parent terminal, the second question search interface including a plurality of second search boxes;
[0659] Obtain the actual question-solving needs of the current user, and determine the second actual search content based on the actual question-solving needs;
[0660] Acquire a second actual search content input in the second search box, and acquire a plurality of matching third questions from the corresponding first preset electronic question bank folder;
[0661] An eighth set number of questions are selected from the plurality of third questions as target questions to form a target electronic homework that matches the actual problem-solving needs of the current students.
[0662] In this solution, students or parents can also select questions from the question bank folder according to their actual needs to form electronic homework. The implementation process is similar to the process of forming electronic homework based on the first question search interface on the teacher side, so it will not be repeated here.
[0663] It should be noted that the question bank, search method, and homework assignment method used by students or parents are slightly different from those used by teachers. For example, students assign questions to themselves, which is more suitable for adult users. For example, college students can prepare for CET-4 and CET-6 exams, bar exams, etc. based on the question bank.
[0664] If the homework assignment is completed by the students alone without the intervention of the teacher, after the students complete the homework they assigned and submit it, the electronic homework will automatically become a corrected homework. The students can check the reference answers corresponding to each question to estimate their own accuracy in answering the questions. This makes it suitable for a wider range of application scenarios and can meet the needs of more users.
[0665] In one feasible solution, the processing method further includes:
[0666] On the student side, a student homework management interface is pre-formed;
[0667] Among them, the student homework management interface includes several homework folders constructed according to different subjects using different preset filing methods;
[0668] Among them, the preset filing methods correspond to filing benchmarks such as unsubmitted, submitted, and revised;
[0669] Specifically, homework can be archived using homework status, which can be divided into unsubmitted homework, submitted homework, and corrected homework; if students see any unsubmitted homework, they can go and complete it and clear the unsubmitted homework.
[0670] During the homework archiving process, the subjects will be automatically divided according to the title labels and class labels, and the homework of the corresponding subject will be automatically archived to the corresponding homework folder; after the electronic homework is archived to the corresponding homework folder, the system will automatically add a label to the electronic homework to automatically mark it as submitted homework; for the homework in the homework folder, the teacher’s corrected homework will also be automatically marked as corrected homework; based on all the homework that should be submitted and the submitted homework of a class, all the unsubmitted homework of the corresponding class will be automatically calculated, and the corresponding special folder will be automatically formed. Once the unsubmitted homework is submitted within the specified time, the homework will be automatically updated to the folder of submitted homework, thereby completing the automatic and efficient management of all homework.
[0671] Of course, you can also archive according to different subjects, different homework release times, etc. The specific archiving method can be selected or switched according to the students' usage preferences to meet the personalized archiving management needs of different students.
[0672] In one feasible solution, a teacher homework management interface is pre-formed in the teacher terminal;
[0673] Among them, the teacher's homework management interface includes the processing status information of each electronic homework.
[0674] In the teacher homework management interface of each teacher, the completion status of the electronic homework published in each batch of each class that the teacher is responsible for can be sorted and displayed in sequence. Specifically, it can be displayed in order from recent to recent according to the homework publication time; for the electronic homework published in the same batch, the specific statistics of the class's electronic homework submission, non-submitted, and corrected status, etc.; of course, other dimensional data can also be displayed as needed to achieve better electronic homework viewing and management as much as possible, which will not be repeated here.
[0675] In one feasible solution, the processing method further includes:
[0676] Based on the active homework submission operation, the target electronic homework submitted by the student is obtained;
[0677] Based on the passive triggering of the homework submission operation, the target electronic homework submitted by the student end is obtained.
[0678] Among them, the passive triggering of homework submission operations includes operations where the student's answering time reaches the fixed answering time in the exam mode, or operations where a third party intervenes to collect the results;
[0679] In this solution, students can actively submit their homework by clicking the Submit button when answering questions; for homework with a timer, the system will automatically submit the corresponding homework when the countdown time is up. Other users using other operating systems can also forcibly mark the homework as submitted. For example, if a student submits the homework late, the teacher will collect the corresponding test paper through pre-set permissions and can obtain the questions that the student has answered so far. Blank records will be uploaded for the parts that the student has not answered, so as to achieve timely and effective collection of each electronic homework, which is convenient for subsequent teaching management such as correction.
[0680] In one feasible solution, the processing method further includes:
[0681] After submitting the homework, based on the homework book in the homework folder selected by the student end, you can view the questions in the electronic homework corresponding to the homework book and the trajectory data in each operation layer of each question;
[0682] In this solution, students can continue working on unsubmitted electronic assignments on the original layer. For submitted assignments, students can view previous layers but cannot modify the original layer. For graded assignments marked as graded by the teacher, students can view both the original layer and the teacher's graded layer, allowing them to continue taking notes or making corrections. If the electronic assignment is graded, the total score and the score for each question will be displayed.
[0683] In one feasible solution, if the user viewing attributes of the electronic assignment have been set in advance, the student's viewing authority for the electronic assignment is controlled according to the user viewing attributes.
[0684] In this solution, the teacher can pre-set whether the student side can view the standard answers to submitted or corrected assignments. Students can view or not view the corresponding standard answers based on the set permissions.
[0685] For submitted assignments, some can be viewed after submission but cannot be modified, according to pre-set settings. For exams, to maintain exam confidentiality, you can pre-set that submissions cannot be viewed. Or, for exam-mode assignments, students can only view them within a set time (e.g., 10 minutes) after submitting them. Once the time limit is reached, students will no longer be able to view the assignment. In other words, by setting access permissions for assignments, you can achieve comprehensive management of electronic assignments.
[0686] In one feasible solution, the processing method further includes:
[0687] For the target electronic homework issued in the same batch of the same class, statistics are kept on the submitted and / or unsubmitted target electronic homework, and the homework submission progress data is obtained through analysis;
[0688] For the submitted target electronic assignments, count the electronic assignments that have been graded and / or not graded, and analyze them to obtain the assignment graded progress data;
[0689] In the homework management interface of the teacher's end, multiple electronic homeworks of the same class are displayed in order, as well as homework submission progress data and / or homework review progress data under each electronic homework;
[0690] In this scenario, a teacher typically teaches multiple classes simultaneously, each with slightly different progress. The homework management system can automatically record, manage, and display the assigned assignments for each class, as well as the current status of each assignment. Each assignment can have statuses such as: pending, published, pending collection, pending grading, grading completed, and explanation completed.
[0691] For example, in order to remind himself that he needs to publish homework on Wednesday, the teacher can first create a new draft, create an electronic homework to be published, and set the scheduled publication time of the test paper. When the publication time approaches, the system will remind the teacher to complete the editing of the homework and complete the publication.
[0692] After the electronic homework is released, the system will display the number of students in the class who should submit their homework, the number of students who have actually submitted their homework, and the progress of the homework submission. If the homework deadline has not yet arrived, the teacher does not need to pay attention to this homework. If the homework deadline has arrived, but many students have not submitted their homework, the teacher can inquire about the situation himself or entrust the class representative to help collect the homework; that is, the class representative's student terminal is set with the authority to collect other homework in the class, and can directly view the homework collection status and urge them to do so, but they will not see the specific answers of other students, so as to achieve the purpose of collecting homework, or send messages through the system one-to-one or one-to-many to urge students who have not submitted their homework to submit their homework in time.
[0693] If the deadline has passed and most of the homework has been collected, the teacher can start grading. For the papers to be graded, the system will display the number of papers that need to be graded and the number that have been graded, indicating the grading progress.
[0694] Teachers can mark the homework that has been explained as explained to complete the entire homework process.
[0695] The summary page on the teacher's side displays the status of all relevant assignments for all current classes that a teacher is responsible for, the tasks that need to be completed, and other information that users need to know about the status of each assignment. This allows teachers to see the latest status of all assignments at a glance, preventing teachers from forgetting to correct or explain the progress, etc., thereby effectively assisting teachers in managing assignments and greatly alleviating the pressure on teachers in managing assignments in actual teaching.
[0696] In one feasible solution, the processing method further includes:
[0697] If the unsubmitted electronic homework is still not submitted within the first preset time period, a first reminder message is generated to remind the student to submit the homework;
[0698] If the time left to the preset deadline of the uncorrected electronic homework is less than a first set value, or the time left to the preset deadline is longer than a second set value, a second reminder message is generated to remind the user to correct the corresponding homework.
[0699] In this solution, if a student has an assignment that has reached the deadline but has not yet been submitted, the system automatically generates a reminder message and pushes it to the student's end, reminding them to complete the assignment and submit it. This reminder message can also be manually posted by the teacher from another end's operating platform. For teachers, if there is an assignment that has not been graded and is far from the release date for the electronic assignment, the system automatically generates a reminder message and pushes it to the teacher's end, reminding them to complete the grading as soon as possible. In other words, by continuously monitoring the student and teacher ends and providing timely reminders when necessary, the rationality and effectiveness of the overall management of electronic assignments are guaranteed.
[0700] In one feasible solution, the processing method further includes:
[0701] In the teacher's end, select the target grading mode in the assignment grading interface to grade multiple electronic assignments in the same batch;
[0702] Among them, the target grading mode includes grading the entire electronic homework of different students one by one until the grading of all students' electronic homework is completed; or batch grading the same question, and processing the questions of the electronic homework one by one until the grading of all students' electronic homework is completed.
[0703] Clicking the mode switch button in the grading mode generates a switching command. For example, command 1 triggers grading for the entire electronic assignment; command 2 triggers batch grading for the same question. Command 1 automatically sorts the grading for the entire electronic assignment by student name (randomly); command 2 automatically extracts the same question from different papers based on the question label for grading. The implementation principle of automatic grading is similar to the previous content and will not be repeated here.
[0704] In this solution, for example, for a certain homework of a certain class, if there are questions 1-10, and there are five students A, B, C, D, and E in the class, when grading from the perspective of different students, the system will display the grading pages of student A's questions 1, 2...10 in sequence for the teacher to perform grading operations, and then display the pages of students B, C, D, and E. When grading from the dimension of each question, the system will display the grading page of question 1 in the homework of students A, B, C, D, and E in sequence, and then display the grading page of question 2 in the homework of students A, B, C, D, and E, and so on, until the batch grading of 10 questions for 5 students is completed. Among them, when grading according to the question dimension, the teacher only needs to remember the content and grading standards of one question at a time, which helps to complete the grading more quickly. For classes with more students, grading according to the question dimension can more effectively improve the efficiency of homework, and effectively relieve the teacher's grading pressure and reduce the time invested in grading.
[0705] In one feasible solution, the processing method further includes:
[0706] Obtain several target electronic assignments of the same batch and belonging to the examination mode submitted by the student end;
[0707] Extracting a set of student answer questions corresponding to one or more questions in a number of target electronic assignments;
[0708] According to the preset allocation rules, the student answer question set is split into multiple sub-answer question sets;
[0709] Send multiple sub-question sets to different teacher terminals for correction until the correction of each target electronic assignment is completed.
[0710] Specifically, if m classes take the joint exam, a1+a2+...+a m Students take the joint entrance examination, and there are n teachers grading them. Enter the username of each teacher, and the system will automatically grade the multiple-choice questions and sort out the distribution range. The remaining questions will be evenly distributed among the teachers based on the principle that the teacher's grading time is proportional to the score of the question. The distribution results are automatically output. After the distribution results are output, manual adjustments can be made, such as selecting p teachers to reduce tasks and q teachers to increase tasks. The system will then redistribute the tasks again to obtain the final distribution results.
[0711] In this solution, in an exam scenario (especially for large-scale exams, such as municipal exams), the number of participating students will be 10,000 or even more. To ensure fairness, teachers will not know which student's answer sheet they are grading during grading, and the exam papers need to be assigned to multiple teachers for collaborative grading. The exam organizer can pre-set relevant grading distribution rules, automatically split the submitted electronic assignments of the same batch into multiple sets of answer questions based on the questions, and then automatically assign the corresponding grading tasks to the relevant responsible teachers for division of labor and grading. Each person completes his or her own grading task, and the grading is completed in batches, thereby ensuring the efficiency and efficiency of the grading of electronic assignments, as well as the accuracy and fairness of the overall grading and scoring of electronic assignments.
[0712] In one feasible solution, the processing method further includes:
[0713] For several target electronic assignments in the same batch after correction, the error rates of the same questions and sub-questions are statistically obtained;
[0714] Each question and sub-question is marked according to the error rate.
[0715] In this solution, the homework management system can also calculate and identify the error rate of each question and sub-question, so that teachers can understand the overall mastery of the corresponding students on which knowledge points, and provide targeted and focused explanations to achieve high-quality auxiliary teaching results.
[0716] In one feasible solution, the processing method further includes:
[0717] Sort and display different questions and sub-questions according to the error rate;
[0718] In this solution, the questions and sub-questions in the electronic homework of the same batch are sorted in descending order according to the error rate, which helps teachers know which questions need to spend more time explaining. The knowledge points or methods involved in the questions with a high error rate require teachers to consolidate their teaching more, so as to achieve more reasonable and effective auxiliary teaching results, further ensuring the overall teaching quality.
[0719] In addition, compared with the traditional paper-and-pencil homework scenario, in this solution, for several electronic homeworks in the same batch, the homework management system automatically completes the data statistics of these electronic homeworks (highest score, lowest score, average score and its corresponding students, etc.) based on the teacher's correction and grading records. The statistical results help teachers better understand the students' mastery of each knowledge point in the entire class, and explain different questions in a targeted manner when explaining the questions.
[0720] In one feasible solution, the processing method further includes:
[0721] On the teacher side, a preset operation is obtained, and based on the preset operation, several student answers for the selected current question or current sub-question in the target electronic homework are obtained;
[0722] Obtain the selected viewing mode, and use the adapted viewing mode to view all incorrect answers and / or all correct answers in several student answers.
[0723] In this solution, if a teacher needs to see the answers of different students to the corresponding questions when preparing lessons or giving lectures, the system can automatically retrieve the answers and scores of different students to the current questions in turn and display them flatly. In this way, the teacher can see how different students made mistakes, such as whether they did not try this question or made a mistake at a certain step.
[0724] As shown in Figure 31, for the specific viewing method, when choosing to view the incorrect answers and correct answers corresponding to all students simultaneously, all deducted answers are sorted before all full-mark answers. For multiple-choice questions, the system can automatically count the number of people who chose each option. Among them, the order of displaying answers can have different modes, such as displaying student answers in the default student order, or placing incorrect answers before correct answers. This can be displayed and explained more intuitively, making it easier for students to discover their own problems and better understand and master the corresponding knowledge points. While optimizing the teaching method, it further ensures the quality of teaching.
[0725] In one feasible solution, the processing method further includes:
[0726] For several target electronic homeworks of the same batch after correction, several student answer contents under the same topic or the same subtopic are obtained according to the actual selection operation to generate several sets of lesson preparation contents.
[0727] In this solution, the homework management system can conduct targeted statistical analysis based on the electronic homework corrected in the same batch, and can also automatically generate a demonstration page for explanation. During the explanation, the user can call up the demonstration page to cooperate with the explanation; specifically, according to the teacher's choice, the question content, standard answers, and some student answers selected by the teacher are displayed on the demonstration page, such as the question content and the answers of three students who did not get full marks on the question; during the explanation, the error methods of each student can be analyzed to guide the students' existing thinking problems, etc., which simplifies many manual operations. The teacher can complete the acquisition of relevant content to be demonstrated by simply checking the box and other operations; preferably, the student's name can be hidden in the demonstration page to protect the student's privacy.
[0728] Specifically, for any question, GPT is used to generate the following contents in sequence and splice them together: the explanation manuscript corresponding to the question; the student questions and answers corresponding to the question; the typical wrong answers corresponding to the question and the explanation of why they are wrong, so as to obtain high-quality explanation content, which is convenient for students to digest and understand; among them, 1-2 most common wrong answers are selected for each question to facilitate understanding, thereby improving the effectiveness of the explanation.
[0729] In one feasible solution, the processing method further includes:
[0730] Get all the corrected historical electronic homework of the same student under the same subject within the set time period before the current time;
[0731] Obtain the initial processing trajectory data of each question in each corrected historical electronic homework by the current student;
[0732] Extracting first processing trajectory data from the initial processing trajectory data; wherein the first processing trajectory data is used to represent trajectory data of the student correcting the deducted points in the corrected electronic homework;
[0733] Based on the first processing trajectory data and the total number of deducted points in the corrected target electronic homework, the homework correction habit data of the current student in the set time period is analyzed and obtained.
[0734] In this solution, all of a student's history homework assignments for a given subject can be regularly aggregated and analyzed to identify the student's corrections to incorrect questions in each assignment. This data can then be used to identify their correction habits. For students with poor correction habits, targeted intervention and guidance can be provided when necessary. Correction behavior statistics can be used to understand students' learning habits and determine whether poor learning outcomes stem from a lack of corrections. This can help students develop better study habits and improve their learning efficiency.
[0735] The analyzed data can also be shared with other users, such as teachers and parents, who can check whether students have corrected any errors in each electronic assignment, and how many corrections have been made. For students with poor correction habits, teachers and parents can initiate reminders through the system or manually remind them to correct their mistakes as soon as possible.
[0736] In addition, the homework management system counts the P questions for which the student did not get full marks within a period of time, and counts how many of these questions the user performed correction operations on, so as to statistically determine the student's homework correction habits during this period.
[0737] For a certain electronic homework, if the user has not corrected the homework after receiving it for a period of time, the system will generate a reminder to remind the user to correct the homework; if necessary, multiple reminders can be set, that is, once every two days, until the user starts to correct the homework, then the reminder will be cancelled.
[0738] Among them, for students' wrong questions, after a certain period of time, it is traced whether the correction is completed or the correction status; for questions that have been corrected, they will be automatically marked as corrected and the corresponding correction data will be automatically saved; in the process of tracing the correction status, it is judged whether the students have corrected each wrong question, and the students' correction completion degree for all wrong questions is known; and generally speaking, the more points lost in a question, the longer it takes to correct it. For questions that students have corrected, it is determined whether they have corrected them seriously based on their loss of points and the correction time. If the correction is sloppy, targeted reminders can be given.
[0739] In one feasible solution, the step of analyzing and obtaining the current student's homework correction habit data within a set time period based on the first processing trajectory data and the total number of deducted questions in the corrected target electronic homework includes:
[0740] For a corrected target electronic homework, if the first processed trajectory data indicates that the current student has triggered the generation of a correction layer and written correction content, or regenerated the handwritten answer area corresponding to the question and written correction content, then it is determined that the current student has corrected the corresponding deducted question, and first confirmation information that the current student has corrected the current handwritten question is generated;
[0741] If the first processed trajectory data indicates that the current student has accessed a draft paper and has draft trajectory data, or has touched a question stem layer and generated annotation data within the question stem area, then it is determined that the current student has corrected the corresponding deducted question, and second confirmation information is generated indicating that the current student has corrected the current question.
[0742] Otherwise, third confirmation information is generated indicating that the current student has not made any corrections to the current deducted point question;
[0743] Obtaining, based on the first determination information, the second determination information, and the third determination information, the number of corrections completed by the current student for the graded electronic assignment, and calculating a ratio of the number of corrections completed to the total number of deducted points for the graded electronic assignment as the current student's correction completion degree for the graded electronic assignment;
[0744] According to the correction completion degree of each corrected homework, the homework correction habit data of the current student in the set time period is obtained.
[0745] In this solution, after students receive the corrected electronic homework, if the user adds an answer area to a question, leaves writing tracks on a new layer after the corrected layer, leaves writing tracks on the draft paper after the homework is corrected, or leaves new circles in the question area, etc., it is considered that the student has corrected the corresponding question. Otherwise, it is considered that no correction operation has been performed. That is, the above-mentioned data changes are generated by automatic analysis, so that it can be timely identified whether the user has made corrections, thereby ensuring the timeliness and accuracy of the analysis of each student's homework correction habits.
[0746] In one feasible solution, the processing method further includes:
[0747] Obtain the grading data of X electronic assignments for each student, or obtain the grading data of Y graded questions for each student; where X and Y are positive integers;
[0748] Obtain the score of each question or sub-question under the question based on the correction data;
[0749] Obtain the answer-related data of each user when answering each question or sub-question;
[0750] Generate a first homework analysis report for each student based on each score and / or answer-related data; and / or generate a second homework analysis report corresponding to one or more classes;
[0751] Among them, the first homework analysis report is any student's mastery of the learning objectives and changes in learning status;
[0752] The second assignment analysis report includes one or more classes’ mastery of learning objectives and changes in learning status.
[0753] Analyze the wrong questions corresponding to the first homework analysis report and the second homework analysis report; or analyze all questions to obtain analysis results;
[0754] In this solution, the homework management system collects raw data about students' learning progress during the homework process, including but not limited to which questions students attempted, the teacher's grading, whether students lost points or how many points they received for each question. By processing this raw data, statistical data for different purposes can be obtained:
[0755] For students, a. After receiving a graded assignment, students can first view a summary report for that assignment when they click on it. This report includes the total score, score percentage, and the correctness of each question. The score for each question can be combined with the question's label information to further assess the student's mastery of the learning objectives. For example, questions contain knowledge point labels, which are a dimension of a student's learning objectives. If questions 1-5 of an assignment have the knowledge point label "addition" and questions 6-10 have the knowledge point label "subtraction," the system will update the student's mastery of both knowledge points based on the scores for questions 1-5 and 6-10, respectively, informing the student, "In this assignment, addition is well mastered, but subtraction needs improvement." Learning objectives can also have other dimensions, such as problem-solving skills, such as combining numbers and shapes, and proof by contradiction. b. The scores of multiple assignments over a period of time can be combined to assess the student's mastery of different knowledge points. c. You can compare the summary reports of the same student at different times or for different assignments to get information about changes in the student's learning, such as "During this period, the addition knowledge has improved, but the subtraction knowledge has regressed and been forgotten, and needs to be consolidated", or compare the student's relative mastery level with other students.
[0756] For a class, a. Statistics for a specific assignment for the entire class can be obtained, including each student's score on this electronic assignment, frequently incorrect questions, highest score, lowest score, score distribution, and which questions they performed well and poorly on. This data can also provide information on the class's overall mastery of a specific learning objective, as well as comparisons with other classes at the same time. This information can help the teacher determine which content needs to be specifically reinforced. b. The overall class scores for multiple assignments over a period of time can be combined to determine the class's mastery of different knowledge points, such as whether the class is better at addition than subtraction. c. Summary analysis reports for the entire class, performed at different times or on different assignments, can be compared to reveal overall learning changes over time, such as "improvement in addition knowledge points during this period." The current class's performance can be compared with other classes to assess relative mastery and trends.
[0757] For a grade, a. you can view the learning progress of the entire grade at a certain stage, and compare it with the learning progress of the same grade in other schools to obtain comparative results. b. You can compare summary analysis reports of the entire grade at different times or different assignments to obtain the overall learning changes of the grade over a period of time, such as "During this period, the knowledge of addition has improved."
[0758] In addition to score information, the above statistical data generally also refers to parameters such as the time students take to answer questions. For example, if a student's accuracy rate for questions on a certain learning objective meets the requirements, but the time spent is significantly longer than the normal time, it will also affect the judgment of the student's mastery of the learning objective. Similarly, it will affect the overall analysis of the entire class and the entire grade.
[0759] In one feasible solution, the corrected questions correspond to label information of several dimensions;
[0760] Get the scores of each question or sub-question in multiple graded electronic assignments of the current student within a set time period;
[0761] Get the answer-related data of the current user when answering each question or sub-question;
[0762] Based on the scores and / or answer-related data, obtain the current student's mastery of the target assessment knowledge points under the statistical dimensions of different label information.
[0763] Among them, the determination of students' progress in mastering knowledge points:
[0764] The knowledge graph, which is the outline of each subject's knowledge points, represents the student's learning goal. The mastery variable x is used to measure the student's mastery of each knowledge point, with x=1 indicating complete mastery and x=0 indicating no mastery at all. Mastery is correlated with the difficulty variable of the question. All questions on the same knowledge point are grouped together, and the score rate for that knowledge point is calculated as the score divided by the total score.
[0765] Count all the questions corresponding to each knowledge point and distribute the difficulty of all questions normally. (y min =μ-2σ, y_max=μ+2σ), get the 5 questions that students have done recently for this knowledge point, the difficulty coefficients are y1, y2, y3, y4, y5, and the mastery level variables are e1, e2, e3, e4, e5, x=(yi-y min -σ) / (y max -y min )*ei average. Mastery level variable e i. Initially set to the percentage of each question's score. If the student's score percentage for the question is above 0.8, but the time spent on the question is at the longer end of the normal distribution, the e value is also lowered. i . The degree of a student's mastery of a certain knowledge point is related to the relative position of the difficulty of the questions on that knowledge point that the student has done among the difficulty of all questions on that knowledge point. If there is only data on students doing simple questions, even if they get all the simple questions right, this calculation method will determine that they have not mastered the knowledge well. For example, for some knowledge points, the difficulty variable is distributed between 0-0.4. For questions with a difficulty of 0.4, if 5 out of 5 questions are correct, x = 1; for some knowledge points, the difficulty variable is distributed between 0.2-1. For questions with a difficulty of 1, if 5 out of 5 questions are correct, x = 1; for some knowledge points, the difficulty variable is distributed between 0.2-1. For questions with a difficulty of 0.2, if 5 out of 5 questions are correct, x is also very low.
[0766] Determination of the student's overall learning situation (combined with learning habits and attitudes):
[0767] Calculate the total weekly study time: Is there a fixed pattern in study time every weekday? If so, is it from 5-6pm or 8-9pm every day?
[0768] Study schedule: Are you used to studying for long periods of time, say an hour at a time, or are you used to studying in segments, say a few times a day, for 20 minutes each time?
[0769] Do you have the habit of correcting wrong answers?
[0770] Regarding the extracurricular knowledge and recommended links recommended in the system, will they click on the additional learning, and what the additional learning situation is like?
[0771] Subject interest: This can be determined by analyzing the students' distribution of study time in each subject, the comparison of their mastery of each subject, the relationship between the extracurricular knowledge recommended in the system, recommended links, click frequency, and subjects.
[0772] In this solution, a student's mastery of different learning objectives is determined based on their scores across all questions attempted over a period of time and associated answer data (such as time spent answering, answering trajectory data, and draft trajectory data). For example, for a mathematics learning objective knowledge point map, each student's mastery of a specific knowledge point within the target knowledge point map can be determined based on their performance, along with a comprehensive overview of their learning situation. This allows students to understand which knowledge points they have mastered and which they have not, allowing them to focus on strengthening their learning of those points. This also facilitates understanding for teachers, allowing them to provide targeted tutoring.
[0773] In one feasible solution, the processing method further includes:
[0774] Get a number of wrong questions from each corrected homework of the current student for the same subject and store them in the wrong question database of the current student;
[0775] According to the preset wrong question training rules, a number of questions are selected from the wrong question bank and form a first electronic homework for wrong question training;
[0776] Alternatively, the wrong question selection parameters input by the student are obtained, and a number of questions are selected from the wrong question bank to form a second electronic homework for wrong question training;
[0777] The wrong question selection parameters include question identification information and / or other preset selection information.
[0778] In this solution, any questions for which students did not receive full marks, i.e., incorrect answers in history, will be automatically added to the corresponding incorrect question bank. Of course, students can also manually add questions for which they received full marks but felt they had not mastered them, such as questions for which they had to guess during the answering process, to their favorite question bank.
[0779] Among them, wrong questions can be classified according to subjects, knowledge points, time of doing questions, whether they have been tried again, etc.; of course, they can also be classified according to the user's own marked label information.
[0780] Students can click to view all the wrong questions in the wrong question bank and select the wrong questions for training, or the homework management system can automatically push wrong questions according to a certain schedule. For example, if a user clicks to practice on wrong questions in math, the system can push wrong questions for different knowledge points in the wrong history questions in math to the student in stages, based on the knowledge points and forgetting curve of each question, so that the student can consolidate and practice the corresponding knowledge points in a timely manner.
[0781] Specifically, the automatic sorting of the wrong answer books can be performed in the following manner (see FIG32 , which shows the display interface of different sorted wrong answer books):
[0782] Method (1): Automatically classify according to knowledge points, specifically classify wrong questions according to other dimensional labels of the system, and conduct special exercises to solve specific mistakes made by students, such as careless mistakes, reading errors, calculation errors, etc.
[0783] Method (2): Obtain the classification method of students who have a high level of mastery of the subject knowledge points and recommend the classification method to other students. The classification method is mainly recommended to students who have significantly improved their progress in mastering the knowledge points over a period of time.
[0784] In addition, the system automatically assigns a review plan for wrong questions, creating a consolidation plan based on the memory curve. The consolidation plan is divided into two parts: reviewing wrong questions and re-doing wrong questions. For example, with the date the correction results for each question are issued as day 0, re-doing wrong questions will be assigned on days 2, 7, and 15, and reviewing wrong questions (only reading the questions and taking notes, no re-doing) will be assigned on day 4.
[0785] It can also automatically calculate the day's tasks based on the student's wrong question records and make corrections based on them. The specific correction methods are as follows:
[0786] Students who have made more mistakes should choose questions with the greatest utility;
[0787] According to the plan of reviewing 1-2 subjects every day, the exercises required for the subject in nearly three days can be completed in one day. Each subject can be reviewed on one day per week;
[0788] Arrange the distribution time of weekly practice questions based on the differences in students' learning and mastery of each subject.
[0789] In one feasible solution, the processing method further includes:
[0790] After the current student processes the questions in the wrong question bank using different question identification information, a new wrong question bank is updated;
[0791] In this solution, the question identification information of any question in the wrong question bank can be modified to obtain an updated wrong question bank to ensure the accuracy of subsequent wrong question training recommendations.
[0792] Treatment methods also include:
[0793] Generate a training set of questions for each student;
[0794] According to the actual homework training needs, several questions are selected from the question training set to form a training homework.
[0795] In this solution, the system automatically selects untried questions from the corresponding subject's question bank folder and recommends them to students for practice. For example, based on the user's mastery of the knowledge point map, questions related to weak points can be selected and recommended. Alternatively, matching questions from the question bank folder or the wrong question bank can be selected and recommended based on the knowledge points corresponding to questions that the student has previously failed, achieving targeted training results.
[0796] In one feasible solution, the steps of generating a training set of questions for each student include:
[0797] Determine the student's learning situation type based on their current understanding of the assessment content;
[0798] Generate a question training set for the current student based on questions that match the learning situation type and questions that match the assessment content with a mastery level lower than the set value;
[0799] In this solution, the question training set that matches the current student can be determined based on the questions corresponding to the student's learning situation type, the questions corresponding to the knowledge points that the student has mastered less, etc.
[0800] In addition, questions can also be recommended based on a preset big data algorithm. Specifically, based on the aggregation of wrong answers from different students, the student's learning type is obtained. Then, questions that students of the same type often get wrong are included in the question training set and recommended to the current student to achieve the effect of targeted training.
[0801] In a feasible solution, a number of wrong questions and corresponding knowledge points of each corrected homework of the current student in the same subject are obtained;
[0802] Based on several wrong questions, several questions that match the knowledge points are obtained from the preset question bank to generate a question training set for the current student.
[0803] In this solution, the knowledge points in which the current student is weak are automatically analyzed, and then the matching related questions are automatically obtained as questions in the question training set and recommended to the current student to achieve the effect of targeted training.
[0804] In one feasible solution, the processing method further includes:
[0805] Obtain several training sample data;
[0806] The training sample data includes student sample feature data and sample label data used to characterize the type of students;
[0807] The characteristic data of each student sample is used as input, the corresponding sample label is used as output, and the preset network model is used for training to obtain the student type prediction model;
[0808] Obtain actual characteristic data of several students;
[0809] Input the actual feature data into the student type prediction model and output the student type of each student;
[0810] Based on the student type to which the current student belongs, a matching first tutoring strategy is generated and sent to the parent end; and / or, a matching second tutoring strategy is generated and sent to the student end.
[0811] Without technological support, teachers typically rely on experience and intuition to categorize students, such as whether they are intelligent, hardworking, or more skilled in liberal arts or science. Big data analysis can reveal previously unknown student attributes and uncover more student categories, providing new insights for educational research and guiding teachers to tailor instruction to individual students, ultimately promoting higher-quality teaching.
[0812] Based on any student's operational data within the system, the system extracts relevant student characteristics and, using a trained predictive model, derives a student type. Educational researchers can pre-determine different tutoring approaches based on these student types and recommend them to students and their parents who meet these characteristics. This approach provides students with appropriate learning methods and parents with appropriate coping strategies.
[0813] Student activity records in the system can be used to generate sample student characteristic data. This data includes scores in different subjects, study habits (homework completion, whether homework is completed on time, whether homework is completed at a fixed time each day), and other usage data (whether students frequently review their work, review past questions and take notes, and whether they frequently check their study reports). Using a large amount of historical data from students across multiple schools and grades, a pre-set network model is trained to cluster several common student categories.
[0814] Among them, the students are classified by clustering:
[0815] Extracting student features: Static features include region, grade, test score records for each subject, mastery of each question, and study time data; Dynamic features include changes in grades and answering speed. For categorical features, such as grade, a one-hot vector is used. Specifically, if there are m grades and a student is in grade i, the vector has m dimensions, with all digits except the i-th digit being 1 being 0. For numerical features, the normalized values are stored. For example, if the score is 60 out of 100, 0.6 is stored to keep it between 0 and 1. These are concatenated to form an n-dimensional vector representing the student features, with n and m being preset values to obtain a student sample with sufficient information for clustering. Due to the potential connections and certain structure between features, the same clusters are not necessarily distributed in Euclidean space. Furthermore, given the large number of students, a density-based clustering algorithm (such as HDBSCAN) is used to achieve good clustering results in non-Euclidean space while reducing the impact of outliers. Furthermore, there is no need to manually select hyperparameters such as the number of clusters. For each class, several student samples at the cluster center were provided to education experts to manually create labels.
[0816] Specifically, the coaching strategy is generated as follows:
[0817] Obtain the parents' anxiety level, parents' understanding of educational methods, etc.;
[0818] Among them, the system is compatible with the chat function, and students and teachers, teachers and parents, etc. can communicate directly based on the chat function; the degree of parental anxiety can be obtained by obtaining the chat content between students and teachers over a period of time and analyzing the text of the conversation content, including analysis of the keywords contained (for example, by pre-setting a large number of keywords such as positivity, negativity, high expectations, confusion, etc., and then extracting and comparing the text in the conversation content, and according to the frequency of occurrence of one or more key words, it is determined whether the parents are in a positive state or an anxious state), or analysis based on the semantic analysis model; in addition, the frequency of practice between parents and teachers can be combined for judgment (generally, the higher the frequency, the higher the possibility of parental anxiety; conversely, the lower the possibility of anxiety); the degree of parents' understanding of educational methods can be determined through self-evaluation, mutual evaluation, questionnaires, etc.
[0819] Label each student in different dimensions: including but not limited to starting level, learning speed, etc.
[0820] Based on the data, we screen students with better learning results and students with similar starting backgrounds;
[0821] Extract good student learning methods, parent education methods, etc., share student cases and parent cases, how parents participate in students' learning, etc., push them to students and parents with similar tags, share learning and achieve common improvement.
[0822] At the same time, we can also filter out students with poor results and similar starting backgrounds from the data, and push them as negative cases to students with similar labels to serve as negative reminders.
[0823] In this solution, a preset model for predicting student types is constructed based on a preset network model, so as to recommend appropriate tutoring strategies to parents and students based on the student type of any student, so as to achieve more accurate teaching and tutoring effects.
[0824] In one feasible solution, the processing method further includes:
[0825] Access the syllabus content structured in the default authority end; the default authority end includes the examination board, etc.
[0826] Among them, the syllabus content corresponds to the assessment scope of the knowledge points;
[0827] When selecting a question from the electronic question bank folder, determine whether the current question is beyond the assessment scope based on the knowledge points of the current question; among them, the target questions are all within the assessment scope;
[0828] Examination boards update the syllabus and exam scope annually, uploading both the old and new versions. By comparing text differences, the boards automatically highlight changes (including deletions and additions). Content in the old lesson plan folder that exceeds the new syllabus is automatically moved to the "Old Syllabus" folder. For added content, teachers can select questions covering the same knowledge point from other examination boards and provide them as a reference for lesson preparation.
[0829] Compare the question-setting styles of different examination bureaus. For example, consider questions covering the same knowledge point in the mathematics editions of the People's Education Press and the Jiangsu Education Press. Teachers will intuitively perceive differences in question-setting styles and key areas of focus. This analysis can be quantified through data analysis. Clustering algorithms can be used to identify consistent examples and identify the dimensions where significant differences exist. Using clustering algorithms, questions from the People's Education Press that are closer to the Jiangsu Education Press style can be identified and recommended to teachers of the Jiangsu Education Press, and vice versa. If teachers or examination bureau researchers wish to understand the question-setting styles and key areas of other examination bureaus and draw on their findings, the system can generate questions from those bureaus for a specific knowledge point and present them to them according to categories such as similar, different, or significantly different styles.
[0830] For the exam administration, based on last year's exam results and student performance, we recommend adjustments to specific areas of assessment. For example, some questions may be targeted by teachers using a "sea of practice" tactic to help students improve their scores, but this doesn't necessarily improve their abilities. These questions are heavily test-oriented, and scores can be improved through extensive practice. We recommend removing questions with these types of assessment methods and styles. Clustering algorithms are used to identify which dimension corresponds to each indicator. Some scores can be increased simply by practicing, while others require student understanding or focus more on assessing student ability.
[0831] In this solution, different examination bureaus, such as the National College Entrance Examination Mathematics Test and the Beijing College Entrance Examination English Test, can set the syllabus content, or the scope of the assessment. The syllabus content consists of knowledge points, indicating what content is required to be covered in the examination paper and which knowledge points will be covered in the questions. After the syllabus content of different examination bureaus is stored electronically, it can be compared to see if any question is included in the syllabus of that examination bureau. During the teaching process, the examination bureau will update the syllabus annually based on student learning. The system can automatically mark any changes in the syllabus, indicating which content has been added or removed. Schools that need to take the exam for the corresponding examination bureau can delete questions that exceed the syllabus from the electronic question bank folder based on the syllabus content. For content newly added to the syllabus, the system can recommend relevant questions for schools to update the question bank content. This avoids unnecessary practice of questions that exceed the syllabus, ensures that questions that exceed the syllabus are promptly removed from electronic assignments in actual teaching scenarios, and also ensures timely training on questions on knowledge points newly added to the syllabus.
[0832] In one feasible solution, the processing method further includes:
[0833] Determine the operational authority of executive entities at different levels within the same school;
[0834] For the same school, the level of authority of the executive body to view statistical data is positively correlated with the level of its hierarchy in the hierarchical structure;
[0835] For different schools, the authority to view statistical data of each school is lower than the authority to view statistical data of the higher-level management department.
[0836] In this solution, school administrators can manage and register students and teachers and assign accounts to each user. Users with different roles in the school have their own accounts and can access content related to them and perform related operations.
[0837] The following are the steps to build a school structure:
[0838] (1) Define how many departments a school has, for example: junior high school, senior high school, and international student department; a school may be composed of different departments; (2) Select a syllabus for each department, for example, the People's Education Press version for the main department and the international student department the international examination syllabus. After the setting is completed, the system can also know which schools teach the same syllabus content, establish a connection between the examination bureau and the school departments, and the examination bureau can manage the school departments and issue notices; schools teaching the same syllabus can also exchange teaching experiences with each other; (3) Determine the number of grades in each department, for example: In-system high schools: Grade 1, Grade 2, Grade 3; International schools: Grade 9, Grade 10, Grade 11, Grade 12; Domestic Chinese university subjects: Grade 1, Grade 2, Grade 3, Grade 4; British university undergraduates: Grade 1, Grade 2, Grade 3. The classification and naming of grades should be selected from the default settings first, but can also be customized. If the school has its own requirements for grade naming, it needs to be aligned with the standard grades after ...
Claims
1. A method for obtaining an answer area, characterized in that: The acquisition method is applied in electronic operations, and the acquisition method includes: Identify the issues to be addressed; Obtaining the topic attribute information of the topic to be processed; Based on the question attribute information, determine whether the question to be processed is a handwritten question type, and if so, generate a handwritten answer area matching the question to be processed in a first preset blank canvas; The first preset blank canvas is a carrier for handwriting input content.
2. The method for obtaining the answer area according to claim 1, characterized in that: The handwriting answer area is used to display the corresponding stacked contents in all operation layers that have been completed and newly generated and are stacked in sequence; Among them, all the operation layers stacked in sequence are generated in sequence when different users enter the handwriting answering area in different preset use stages.
3. The method for obtaining the answer area according to claim 2, characterized in that: The operation layers generated in different preset use stages include an initial layer for use by a student terminal, a correction layer for use by a teacher terminal, a revision layer for use by the student terminal, or a revision layer for use by the teacher terminal; Wherein, the first preset blank canvas is the initial layer generated based on a canvas in a picture format; After N correction operations, the correction layer for the student end and the correction layer for the teacher end are generated again in the handwritten answer area in sequence, and N is a positive integer.
4. The method for obtaining the answer area according to claim 2, characterized in that: The acquisition method further includes: Obtain a re-answer instruction corresponding to the pending question; Based on the re-answer instruction, a new handwriting answer area matching the question to be processed is generated in a second preset blank canvas; Control all the operation layers in the handwriting answering area obtained based on the first preset blank canvas to be in a display state or a hidden state.
5. The method for obtaining the answer area according to at least one of claims 1 to 4, characterized in that: The handwriting answering area is an area used to respond to and display handwriting touch content; and / or, an area for responding to and displaying content input based on an external input device; The handwriting answer area is displayed in the display interface of the smart terminal that supports handwriting touch operation.
6. The method for obtaining the answer area according to claim 5, characterized in that: The electronic homework includes a plurality of stems and a plurality of to-be-processed questions under each stem; The step of generating a handwritten answer area matching the question to be processed in the first preset blank canvas comprises: For each of the pending questions belonging to the handwriting question type in the electronic homework, the handwriting answer area of each of the pending questions is generated on the first preset blank canvas.
7. The method for obtaining the answer area according to claim 6, characterized in that: The to-be-processed topic includes several topic levels, and each of the topic levels corresponds to several sub-topics; The step of generating a handwritten answer area matching the question to be processed in the first preset blank canvas comprises: When it is determined that the sub-topic belongs to the handwriting question type, the handwriting answer area of the sub-topic is generated on the first preset blank canvas to obtain a plurality of handwriting answer areas corresponding to the question to be processed.
8. The method for obtaining the answer area according to claim 7, characterized in that: The step of generating a handwritten answer area matching the question to be processed in the first preset blank canvas comprises: When there are multiple sub-topics belonging to the handwriting question type, the handwriting answering areas of all the multiple sub-topics belonging to the handwriting question type are generated synchronously; The handwriting answer area of each sub-topic is displayed following the location of the corresponding sub-topic; or, the handwriting answer areas of multiple sub-topics are displayed sequentially and centrally based on the order of the sub-topics.
9. The method for obtaining the answer area according to at least one of claims 1 to 8, characterized in that: After the step of generating a handwritten answer area matching the question to be processed in the first preset blank canvas, the method further includes: Receive region size adjustment instructions; Based on the area size adjustment instruction, the size of the handwriting answer area is adjusted.
10. The method for obtaining the answer area according to claim 9, characterized in that: A first size adjustment plug-in is correspondingly arranged at a first preset position of the handwriting answer area; The step of receiving the region size adjustment instruction comprises: In response to the first size adjustment plug-in, obtaining a region size expansion instruction; The step of adjusting the size of the handwriting answer area based on the area size adjustment instruction includes: Increasing the size of the handwriting answer area to a first preset size based on the area size expansion instruction; and / or, A second size adjustment plug-in is correspondingly arranged at a second preset position of the handwriting answer area; The step of receiving the region size adjustment instruction comprises: In response to the second size adjustment plug-in, obtaining a region size reduction instruction; The step of adjusting the size of the handwriting answer area based on the area size adjustment instruction includes: Reducing the size of the handwritten answer area to a second preset size based on the area size reduction instruction; and / or, A third size adjustment plug-in is correspondingly arranged at the third preset position of the handwriting answer area; The step of receiving the region size adjustment instruction comprises: In response to the third size adjustment plug-in, displaying a parameter input interface; Receiving externally input area size parameters based on the parameter input interface; The step of adjusting the size of the handwriting answer area based on the area size adjustment instruction includes: Adjusting the size of the handwritten answer area to a third preset size based on the area size parameter; And / or, the default size of the handwritten answer area is positively correlated with the score of the corresponding question to be processed.
11. The method for obtaining the answer area according to at least one of claims 1 to 10, characterized in that: The handwriting answering area supports obtaining writing track related data; Wherein, the writing track associated data is stored in a preset storage space; The writing trajectory associated data includes a plurality of writing trajectory data segments, and first time frame data and / or answer writing speed data corresponding to each writing trajectory data segment; In the student terminal, the writing track data is the answer track data; In the teacher's end, the writing track data is the answer correction data.
12. The method for obtaining the answer area according to claim 11, characterized in that: The answer box area supports dynamic following display of the writing trajectory associated data in the target end.
13. The method for obtaining the answer area according to claim 11 or 12, characterized in that: The handwriting answering area supports displaying the operation layer of the student terminal in the first preset user terminal; And / or, the handwritten answer area supports replaying the answer track data stored in the preset storage space in a second preset user terminal; And / or, the handwriting answer area supports displaying the operation layer of the teacher terminal in a third preset user terminal; And / or, the handwritten answer area supports playback of the answer correction data stored in the preset storage space in a fourth preset user terminal; Wherein, the corresponding data playback mode is determined based on different selection instructions.
14. The method for obtaining the answer area according to claim 11, characterized in that: The handwriting answering area is provided with a corresponding basic function component, and the basic function component includes a plurality of basic function plug-ins.
15. The method for obtaining the answer area according to claim 14, characterized in that: The basic function plug-in includes a draft paper calling plug-in, and the draft paper calling plug-in is used to call or hide the draft paper; The draft paper area where the draft paper is located is generated following the corresponding handwritten answer area, and each of the The handwritten answer area corresponds to one piece of the draft paper; or, all the handwritten answer areas under each of the pending questions correspond to one piece of the draft paper; The draft paper is dynamically stored in a preset storage space of the intelligent terminal.
16. The method for obtaining the answer area according to claim 15, characterized in that: The attribute information of the draft paper is adjustable; And / or, the display style of the draft paper area where the draft paper is located is different from the display style of the handwritten answer area; and / or, when calling the draft paper, determining whether the eraser function of the handwritten answer area is in a disabled state based on the actual selection operation; And / or, the draft paper area is located on the first preset blank canvas where the handwritten answer area is located; And / or, the draft paper area is displayed in a front position or a rear position of the handwritten answer area; And / or, the draft paper area and the handwritten answer area are displayed side by side in different areas.
17. The method for obtaining the answer area according to claim 15 or 16, characterized in that: The draft paper area supports obtaining draft trajectory associated data; Wherein, the draft trajectory associated data is stored in a preset storage space; The draft trajectory associated data includes a plurality of draft trajectory data segments, and second time frame data and / or draft writing speed data corresponding to each draft trajectory data segment; In the student terminal, the draft track associated data is the draft writing track data; In the teacher's end, the draft track associated data is the draft correction track data.
18. The method for obtaining the answer area according to claim 17, characterized in that: The answer box area supports dynamic following display of the draft track associated data of the draft paper in the target end.
19. The method for obtaining the answer area according to claim 17, wherein: The draft paper area supports displaying the draft writing track data of the student terminal in the first preset user terminal; And / or, the draft paper area supports playback of the draft writing trajectory data stored in the preset storage space in a second preset user terminal; And / or, the draft paper area supports displaying the draft correction track data of the teacher terminal in a third preset user terminal; And / or, the draft paper area supports playback of the draft correction track data stored in the preset storage space in a fourth preset user terminal; Wherein, the corresponding data playback mode is determined based on different selection instructions.
20. The method for obtaining the answer area according to claim 17, wherein: The acquisition method further includes: Acquire actual writing data in the answering area for dynamic following display; Wherein, the actual writing data includes the writing track associated data, or is based on the writing track associated data and the draft track associated data; Analyzing the actual writing data using a preset abnormal answer monitoring rule to obtain a first analysis result; Based on the first analysis result, determine whether there is any abnormality in the answering behavior or the correcting behavior. If so, generate and issue a first alarm message.
21. The method for obtaining the answer area according to at least one of claims 11 to 20, characterized in that: The handwriting answering area is provided with an audio interaction plug-in; The acquisition method further includes: In response to the audio interaction plug-in, acquiring and storing externally input audio content and audio-related data; and / or, acquiring and storing externally input audio content and audio-related data, and synchronizing the audio content and the writing track-related data belonging to the same time frame based on the audio-related data, so as to merge the audio content and the writing track-related data to obtain merged data; The handwritten answer area supports playback of the merged data in the handwritten answer area on the target end.
22. A method for processing electronic work, characterized in that: The processing method is implemented according to the method for obtaining the answer area according to any one of claims 1 to 21, and the processing method includes: Based on the preset learning objectives, select several questions from the preset question bank; The acquisition method is used to generate a handwriting answer area corresponding to each question of the handwriting question type; Generate set answer positions corresponding to other remaining questions of non-handwritten question types; The electronic homework is obtained based on the question and the corresponding handwritten answer area and the set answer position.
23. The electronic operation processing method according to claim 22, characterized in that: The handwriting answering area supports obtaining writing trajectory associated data, wherein the writing trajectory associated data includes a plurality of writing trajectory data segments and first time frame data corresponding to each writing trajectory data segment; The processing method also includes: Acquire the dwell time data of each question in the electronic homework at the teacher terminal and / or the student terminal, and the first time frame data of the handwritten answer area; Analyzing and processing the residence time data and the first time frame data to obtain a second analysis result; The marking situation and / or answering situation of the teacher end and / or the student end are evaluated according to the second analysis result.
24. The electronic operation processing method according to claim 23, characterized in that: The processing method also includes: Counting the dwell time data of the same topic in the electronic homework each time, so as to obtain the first dwell time corresponding to the same topic; Obtaining the sum of the first dwell time corresponding to each of the questions in the electronic homework, so as to obtain the second dwell time corresponding to all the questions; Performing analysis and processing based on the first stay duration and the second stay duration to obtain a third analysis result; The time consumption of marking and / or answering questions at the teacher end and / or the student end is determined based on the third analysis result.
25. The electronic operation processing method according to claim 23, characterized in that: The processing method also includes: Counting the dwell time data of the same topic in the electronic homework each time, so as to obtain the first dwell time corresponding to the same topic; Acquire the first time frame data corresponding to the writing trajectory data in the handwriting answer area of the question; Performing analysis and processing based on the first time frame data and the first stay duration to obtain a fourth analysis result; Based on the fourth analysis result, determine whether there is any abnormality in the answering behavior or the correcting behavior. If so, generate and issue a corresponding second warning message.
26. The method for processing electronic work according to at least one of claims 22 to 25, characterized in that: The processing method also includes: Obtaining the answer progress data corresponding to the question and sending it to a preset storage space; The answering progress data includes first characterization data of whether the student terminal has entered the question, and answering track data of the student terminal in the answering area of the question and / or draft writing track data of the draft paper; When the first representation data represents that the student terminal has entered the topic, controlling the first current display interface of the student terminal to be shared and displayed in the first preset user terminal; After the student terminal submits the question, controlling the first preset user terminal to replay the answer track data of the student terminal and / or the draft writing track data of the draft paper; or, Obtaining the correction progress data corresponding to the question and sending it to the preset storage space; The correction progress data includes second representation data of whether the teacher terminal enters the question, and correction track data of the teacher terminal on the answer area of the question and / or correction track data of the draft on the draft paper; When the second representation data represents that the teacher terminal has entered the topic, controlling the second current display interface of the teacher terminal to be displayed on a second preset user terminal; After the teacher terminal completes the question marking, the marking track data of the teacher terminal and / or the draft marking track data of the draft paper are controlled to be replayed in the second preset user terminal.
27. The method for processing electronic work according to at least one of claims 22 to 25, characterized in that: The processing method also includes: For the multiple electronic jobs in the same batch, at least one of the following operations is performed: Obtaining the completion progress of each electronic homework, dynamically tracking and displaying each electronic homework, obtaining each electronic homework for static display, and obtaining each electronic homework for playback; and / or, Control the static display or playback of different handwritten answer contents corresponding to the same handwritten question in different electronic assignments.
28. The electronic operation processing method according to claim 26, characterized in that: The processing method also includes: Determining a calling mode of the draft paper in the electronic homework based on the selection instruction; The calling mode corresponds to calling a common draft paper for use by a plurality of the questions; The public draft paper is reusable; And / or, calling independent draft paper for use with the corresponding topic, each topic corresponds to one independent draft paper.
29. The electronic operation processing method according to claim 28, characterized in that: The processing method also includes: Acquire the annotation data generated when the teacher terminal and / or the student terminal marks the location area of the question stem of the question by handwriting touch; The layer where the annotated data is located is stacked on the layer where the question stem is located; And / or, the annotated data is stored in a preset storage space.
30. The method for processing electronic work according to at least one of claims 22 to 29, characterized in that: The processing method also includes: Build the topic data architecture for topic construction; The topic data structure corresponds to a number of topic levels, and each of the topic levels corresponds to a number of sub-topics; Different subtopic identification data and subtopic label data corresponding to different subtopics, as well as subordinate association data of the topic level to which each subtopic belongs, are obtained and stored in a storage mapping data table.
31. The electronic operation processing method according to claim 30, characterized in that: The processing method also includes: Obtaining a number of selected subtopics under the topic data structure; Based on the external selection instruction, determine the target subtopic to be deleted, so as to update and obtain a plurality of new subtopics selected under the topic data architecture; And / or, the number of topic levels is expandable.
32. The electronic operation processing method according to claim 30, characterized in that: Before the step of constructing the topic data architecture for topic building, it also includes: Using preset decomposition and reconstruction rules to process preset text content to construct a storage data mapping relationship table for multi-dimensional analysis of the topic data architecture; The preset decomposition and reconstruction rules are adapted to the scenario requirements under a number of different teaching specifications; The steps of constructing a topic data architecture for topic building include: The topic data architecture used for constructing the topic is generated according to the storage data mapping relationship table.
33. The method for processing electronic work according to at least one of claims 30 to 32, characterized in that: The processing method also includes: When constructing the topic data architecture or after obtaining the electronic homework, obtaining a topic tag editing operation performed by a user; In response to the topic tag editing operation, the topic tag data or sub-topic tag data of the topic is changed.
34. The electronic operation processing method according to claim 31, characterized in that: The processing method also includes: Preset reference answers corresponding to the sub-questions; The preset reference answer includes a plurality of sub-scoring points, each of which corresponds to a scoring content and a sub-score, and the sum of the plurality of sub-scores is the total score of the sub-question; Obtain the student's handwritten answer corresponding to the answer writing track data in the handwritten answer area; identifying the actual answer points in the student's handwritten answer; Comparing and obtaining the degree of coincidence between the actual answer points and the preset sub-score points; Obtaining the total score of the sub-questions and the actual sub-score of each actual answer point based on the overlap; And / or, the plurality of preset sub-scoring points include a first number of substantial sub-scoring points and a second number of insubstantial sub-scoring points; Wherein, the substantive sub-scoring points and the non-substantive sub-scoring points are arranged according to a preset order; The non-essential sub-score points are marked with zero points or no points.
35. The electronic operation processing method according to claim 34, characterized in that: The processing method also includes: Determine whether the answer viewing condition is met at the current moment, and if so, obtain the student terminal's answer display progress requirement for the sub-question; According to the answer display progress requirements, partial contents of the preset reference answers corresponding to the sub-questions are displayed step by step until the complete preset reference answers are displayed.
36. The electronic operation processing method according to claim 22, characterized in that: The processing method also includes: Generate a preset layout template for the topic based on a physical assignment style or an electronic document assignment style; Wherein, the preset layout template includes a plurality of editable areas associated with different subtopics; The processing method also includes: When a first click operation is obtained at the editable area, a corresponding edit box and an edit tool are displayed; Wherein, when the editable area is in a default state, or when a second click operation is performed on other areas outside the current editable area, the edit box and the editing tool of the editable area are hidden.
37. The electronic operation processing method according to claim 22, characterized in that: The processing method also includes: Get the external input question image; Performing recognition processing on the title image to extract input content; The input content includes the question to be input and / or the answer corresponding to the question to be input; The corresponding sub-topic or topic is generated based on the input content and stored in a preset topic bank.
38. The method for processing electronic tasks according to claim 37, wherein: The input content includes at least one of text content, picture content, and code content corresponding to the calculation formula in the external input image; and / or, The title image is generated based on handwritten content, printed content or electronic content; The handwritten content is content written on paper materials or electronic handwritten pages.
39. The method for processing electronic tasks according to claim 37 or 38, characterized in that: Before the step of selecting a number of questions from a preset question bank based on the preset learning objectives, the method further includes: Constructing a number of said topics; Receive a check result from the user on whether there is a preset error in the question; When the checking result indicates that the preset error exists, a prompt message is generated to prompt the user to correct the current question until the question without the preset error is obtained; When the inspection result indicates that the preset error does not exist, the corresponding question is stored in the preset question library.
40. The method for processing an electronic job according to at least one of claims 22 to 39, characterized in that: The processing method also includes: Pre-build several electronic question bank folders; The same electronic question bank folder corresponds to a number of questions corresponding to the same subject in the same grade; Get the teaching related parameters of each subject in each class of the same grade; Generate a number of initial electronic homeworks corresponding to each class according to the teaching association parameters and the questions in the electronic question bank folder of the corresponding subject; Wherein, several copies of the initial electronic homework for each class are stored in the same homework folder, and different initial electronic homeworks correspond to different preset teaching progresses; The target electronic homework to be assigned each time is determined based on the several copies of the initial electronic homework, and is published to the student terminal.
41. The electronic operation processing method according to claim 40, characterized in that: The processing method also includes: Preset the first question search interface in the teacher's terminal; The first search interface includes a plurality of first search boxes, and the search contents in different first search boxes correspond to search parameters of different dimensions; or, the search contents in the first search box include a preset part of the question stem content or a preset part of the answer content; The step of generating a plurality of initial electronic assignments corresponding to each class according to the teaching association parameters and the questions in the electronic question bank folder of the corresponding subject includes: According to any of the preset teaching progresses, determining a first actual search content input into the question search interface to obtain a plurality of matching first questions from the electronic question bank folder of the corresponding subject; A first set number of target topics are selected based on the first topics to form an initial electronic homework for the corresponding class, until a plurality of initial electronic homeworks corresponding to different preset teaching progresses are obtained.
42. The method for processing electronic tasks according to claim 40, wherein: Before the step of determining the target electronic homework to be assigned each time based on the plurality of initial electronic homeworks and publishing the same to the student terminal, the step further includes: Selecting the first class in the same grade in previous years, whose correlation with the teaching correlation parameter of the current class is greater than a first set threshold; Retrieve the historical related homework of the same teaching progress that has been published under the first class; Using the historically associated assignment as the initial electronic assignment for the current class; or, Selecting a second set number of topics from the historical related homework as target topics, and selecting a third set number of topics from the first topics searched through the first search box as the target topics, so as to obtain the initial electronic homework for the corresponding class; or, Selecting a second class in the same grade of the current year, whose correlation with the teaching correlation parameter of the current class is greater than a second set threshold; Retrieving the first electronic homework with the same teaching progress in the homework folder of the second class; Using the first electronic assignment as the initial electronic assignment for the current class; or, A fourth set number of questions are selected from the first electronic homework, and a fifth set number of questions are selected from the first questions obtained by searching the search box, and both are used as target questions to obtain the said initial electronic assignment for the corresponding class; or, A sixth set number of questions are selected from the plurality of first electronic assignments as target questions to obtain the initial electronic assignment for the corresponding class.
43. The electronic operation processing method according to claim 41, characterized in that: The search interface includes a topic overview area and a topic details preview area; After the step of obtaining a plurality of matching first questions from the electronic question bank folder of the corresponding subject, the method further includes: Control to display in order in the topic overview area a list of topics corresponding to the first topic; The step of selecting a first set number of target topics based on the first topics to form an initial electronic homework for the corresponding class includes: When one or more of the first topics are selected, controlling the display of the content of the selected first topics in the topic detail preview area, and determining whether to use the selected first topics as the target topics according to the actual selection operation, until the first set number of the target topics are obtained, so as to form the initial electronic homework of the corresponding class; and / or, The step of selecting a first set number of target topics based on the first topics to form an initial electronic homework for the corresponding class includes: Selecting the first set number of the target topics based on the first topics, and generating the initial electronic homework for the corresponding class in a physical homework style or an electronic document homework style; and / or, According to the first question in the question details preview area, summarizing and displaying the number of questions and the corresponding total scores of the questions to determine whether the homework formation requirements are met; and / or, Perform at least one of the following operations on different first topics in the topic details preview area: Adjust the order of several first topics, edit the related content of any one of the first topics, delete one or more of the first topics, and mark the topic numbers of several first topics in sequence; and / or, Based on the selected display mode switching instruction, the title contents of the plurality of first titles in the title detail preview area are displayed in a corresponding display mode.
44. The electronic operation processing method according to claim 41, characterized in that: The processing method also includes: Get the actual user operation on the first search interface: Based on the actual operation of the user, obtaining multi-dimensional identification information of the corresponding first question; Determining whether to use the current first topic as the target topic according to the identification information; After the step of generating the initial electronic homework of the corresponding class, the method further includes: Constructing identification parameters of the initial electronic assignment according to the identification information of each target topic; When the identification parameter includes an assignment mode and the assignment mode is an examination mode, identifying the sub-score of each target question, and the total score of the entire initial electronic assignment and / or the corresponding fixed answering time; The fixed answering time is used to provide a countdown reminder in the exam mode, and trigger the submission of the corresponding electronic homework when the countdown reaches zero; When the identification parameter includes an assignment mode and the assignment mode is an after-school assignment mode, no score identification is performed on the initial electronic assignment and / or a corresponding recommended answering time is not determined.
45. The method for processing electronic tasks according to claim 44, wherein: The processing method also includes: When the first topic is determined as the target topic, and the identification information of the first topic indicates that the first topic has been selected and formed into an electronic homework and published to the current class within a set time period from the current moment, a prompt message is generated; and / or, The processing method also includes: The newly added questions are stored in the electronic question bank folder of the corresponding subject, or in a temporary electronic question bank folder.
46. The method for processing an electronic job according to at least one of claims 40 to 45, characterized in that: The step of determining the target electronic job to be arranged each time based on the plurality of initial electronic jobs comprises: Get the actual teaching progress of the current subject in each class; Comparing the actual teaching progress with the preset teaching progress of the initial electronic homework of the corresponding subject to obtain a comparison result; According to the comparison result, the designated topic or the subtopic under the designated topic in the initial electronic homework is adjusted to form a target electronic homework that matches the actual teaching progress of each of the classes.
47. The method for processing electronic tasks according to claim 46, wherein: The step of adjusting the designated topic or the subtopic under the designated topic in the initial electronic homework according to the comparison result to form a target electronic homework matching the actual teaching progress of each class includes: If the comparison result indicates that the actual teaching progress is slower than the preset teaching progress included in the lesson preparation content, the designated questions or sub-questions under the designated questions corresponding to the unfinished knowledge points in the preset teaching progress are deleted from the initial electronic homework to form the target electronic homework that matches the actual teaching progress of each class; or, after deleting questions from the initial electronic homework, according to external requirements for adding questions, the target electronic homework is added to the initial electronic homework. Adding corresponding questions to the homework to form the target electronic homework that matches the actual teaching progress of each class; If the comparison result indicates that the actual teaching progress is faster than the preset teaching progress contained in the lesson preparation content, the designated questions or sub-questions under the designated questions corresponding to the knowledge points that exceed the preset teaching progress will be added to the initial electronic homework to form the target electronic homework that matches the actual teaching progress of each class.
48. The method for processing electronic tasks according to claim 47, wherein: The processing method also includes: Before publishing the target electronic homework on the teacher's end, previewing or editing the target electronic homework; and / or, The step of publishing the target electronic homework to the student terminal includes: Before the teacher terminal releases the target electronic homework and sets the release time in advance, the system time is monitored, and when the system time reaches the release time, the target electronic homework is released to the student terminals of all students in the current class.
49. The method for processing an electronic job according to at least one of claims 41 to 48, characterized in that: The processing method also includes: Obtain the actual learning information of any subject of the target student; According to the actual learning information, determine a second actual search parameter input into the first search box in the question search interface, and obtain a plurality of matching second questions from the electronic question bank folder; Selecting a seventh set number of topics from the plurality of the second topics as the target topics to form the target electronic homework that matches the actual homework needs of the target student for the corresponding subject; and / or, The processing method also includes: Constructing a second question search interface in the student terminal or the parent terminal, wherein the second question search interface includes a plurality of second search boxes; Acquire the actual question-solving needs of the current user, and determine the second actual search content based on the actual question-solving needs; Acquire the second actual search content input in the second search box, and acquire a plurality of matching third questions from the corresponding first preset electronic question bank folder; An eighth set number of questions are selected from the plurality of third questions as the target questions to form the target electronic homework that matches the actual question-solving needs of the current student.
50. The method for processing electronic tasks according to claim 40 or 49, characterized in that: The processing method also includes: In the student end, a student homework management interface is preformed; The student homework management interface includes different preset archiving methods according to different subjects. Several job folders; Wherein, the preset filing mode corresponds to at least one of the filing criteria of unsubmitted, submitted, and revised; and / or, In the teacher's end, a teacher's homework management interface is pre-formed; Among them, the teacher homework management interface includes processing status information of each electronic homework.
51. The method for processing electronic tasks according to claim 50, wherein: The processing method also includes: Based on the active homework submission operation, obtaining the target electronic homework submitted by the student terminal; Based on the passive triggering of the homework submission operation, obtaining the target electronic homework submitted by the student terminal; The passive triggering operation of submitting homework includes the operation that the answering time of the student terminal reaches the fixed answering time in the examination mode, or the operation that a third terminal intervenes and collects the answer; and / or, After submitting the homework, based on the homework book in the homework folder selected by the student terminal, the questions in the electronic homework corresponding to the homework book and the track data in each of all the operation layers in each of the questions are viewed; and / or, If the user viewing attribute of the electronic homework has been set in advance, the viewing authority of the student terminal to the electronic homework is controlled according to the user viewing attribute.
52. The method for processing electronic tasks as claimed in claim 51, characterized in that: The processing method also includes: For the target electronic homeworks issued in the same batch of the same class, statistics are collected on the target electronic homeworks that have been submitted and / or not submitted, and the homework submission progress data is obtained through analysis; For the submitted target electronic homework, counting the electronic homework that has been graded and / or not graded, and analyzing to obtain homework graded progress data; Wherein, in the homework management interface of the teacher's end, multiple electronic homeworks of the same class are displayed in order, as well as homework submission progress data and / or homework correction progress data under each electronic homework; and / or, The processing method also includes: If the unsubmitted electronic homework is still not submitted within the first preset time period, a first reminder message is generated to remind the student to submit the homework; If the uncorrected electronic homework is less than a first set value from the preset deadline, or has exceeded a second set value from the preset deadline, a second reminder message is generated to remind the user to correct the corresponding homework.
53. The method for processing electronic tasks as claimed in claim 52, characterized in that: The processing method also includes: In the teacher's homework grading interface, select the target grading mode to grade multiple electronic assignments in the same batch. Correction; The target grading mode includes grading the entire electronic homework of different students one by one until all students' electronic homework are graded; or batch grading the same question and processing the questions of the electronic homework one by one until all students' electronic homework are graded.
54. The method for processing electronic tasks as claimed in claim 53, characterized in that: The processing method also includes: Obtaining a plurality of target electronic assignments of the same batch and belonging to the examination mode submitted by the student end; Extracting a set of student answer questions corresponding to one or more questions in a plurality of target electronic assignments; According to a preset allocation rule, the student answering question set is split into a plurality of sub-answering question sets; The plurality of sub-question sets are respectively sent to different teacher terminals for marking until the marking of each target electronic homework is completed.
55. The method for processing electronic tasks according to claim 54, wherein: The processing method also includes: For a plurality of the target electronic assignments in the same batch after correction, the error rates of the same question and the same sub-question are obtained by statistics; Marking each of the questions and sub-questions according to the error rate; and / or; According to the error rate, different questions and sub-questions are sorted and displayed; and / or, In the teacher's terminal, a preset operation is obtained, and according to the preset operation, several student answer contents of the selected current question or current sub-question in the target electronic homework are obtained; Obtain the selected viewing mode, and use an adapted viewing mode to view all incorrect answers and / or all correct answers in the plurality of student answers. and / or, The processing method also includes: For the target electronic homeworks of the same batch after correction, the student answer contents under the same topic or sub-topic are obtained according to the actual selection operation to generate several sets of lesson preparation contents.
56. The method for processing electronic tasks as claimed in claim 54, characterized in that: The processing method also includes: Get all the historical electronic homework that has been graded for the same student and the same subject within the set time period before the current time; Obtaining the initial processing track data of each question in each of the corrected historical electronic homeworks by the current student; Extracting first processing trajectory data from the initial processing trajectory data; wherein the first processing trajectory data is used to represent trajectory data of students correcting deducted points in the corrected electronic homework; Based on the first processing trajectory data and the total number of deducted questions in the corrected target electronic homework, the homework correction habit data of the current student in the set time period is analyzed and obtained.
57. The method for processing electronic tasks according to claim 56, wherein: The step of analyzing and obtaining the homework correction habit data of the current student within the set time period based on the first processing trajectory data and the total number of deducted questions in the corrected target electronic homework includes: For a corrected target electronic homework, if the first processing trajectory data indicates that the current student triggers the generation of a correction layer and writes correction content, or regenerates the handwritten answer area corresponding to the question and writes correction content, it is determined that the current student has corrected the corresponding deducted question, and first confirmation information that the current student has corrected the current handwritten question is generated; If the first processing trajectory data indicates that the current student has called up the draft paper and has the draft trajectory data, or has touched the question stem layer and has generated annotation data in the area where the question stem is located, it is determined that the current student has corrected the corresponding deducted question, and second confirmation information that the current student has corrected the current question is generated; Otherwise, third confirmation information is generated indicating that the current student has not made any corrections to the current deducted point question; According to the first determination information, the second determination information, and the third determination information, the number of corrections completed by the current student on a corrected electronic homework is obtained, and a ratio of the number of corrections completed to the total number of deducted points on the corrected electronic homework is calculated as the correction completion degree of the current student on the corrected electronic homework; According to the correction completion degree of each corrected homework, the homework correction habit data of the current student within the set time period is obtained.
58. The method for processing electronic tasks according to claim 54, wherein: The processing method also includes: Obtain the correction data of X electronic homeworks of each student, or obtain the correction data of Y corrected questions of each student; wherein X and Y are positive integers; Obtaining the score of each question or sub-question under the question according to the grading data; Obtain the answer-related data of each user when answering each question or sub-question; Generate a first homework analysis report for each student according to each of the scores and / or the answer-related data; and / or generate a second homework analysis report corresponding to one or more classes; The first homework analysis report is the mastery of the learning objectives and the changes in learning status of any student; The second homework analysis report includes one or more classes' mastery of learning objectives and changes in learning status.
59. The method for processing an electronic job according to at least one of claims 50 to 58, characterized in that: The processing method also includes: Analyze the wrong questions corresponding to the first homework analysis report and the second homework analysis report; or There are questions to analyze in order to obtain analysis results; and / or, The corrected questions correspond to label information of several dimensions; Get the scores of each question or sub-question in multiple graded electronic assignments of the current student within a set time period; Get the answer-related data of the current user when answering each question or sub-question; According to the score situation and / or the answer-related data, the current student's mastery of the target assessment knowledge points under different statistical dimensions of the label information is obtained.
60. The method for processing electronic tasks according to claim 54, wherein: The processing method also includes: Obtaining a number of wrong questions of each corrected homework of the same subject of the current student, and storing them in a wrong question database of the current student; According to preset wrong question training rules, a number of questions are selected from the wrong question library and form a first electronic homework for wrong question training; Or, obtaining wrong question selection parameters input by the student, selecting a number of questions from the wrong question library and forming a second electronic homework for wrong question training; The wrong question selection parameters include question identification information and / or other preset selection information.
61. The method for processing electronic tasks according to claim 60, wherein: The processing method also includes: After the current student processes the questions in the wrong question bank using different question identification information, a new wrong question bank is updated; and / or, The processing method also includes: Generate a training set of questions for each student; According to actual homework training needs, several questions are selected from the question training set to form a training homework.
62. The method for processing an electronic operation as claimed in claim 61, characterized in that: The step of generating a training set of questions for each student includes: Determine the learning situation type according to the current student's mastery of the assessment content; Generate the question training set for the current student based on questions matching the learning situation type and questions matching the assessment content whose mastery level is less than a set value; or, Obtaining a number of wrong questions and corresponding knowledge points of each corrected homework of the current student for the same subject; According to the wrong questions, a number of questions matching the knowledge points are obtained from a preset question bank to generate the question training set for the current student.
63. The method for processing an electronic job according to at least one of claims 40 to 62, characterized in that: The processing method also includes: Obtain several training sample data; The training sample data includes student sample feature data and sample label data used to characterize the type of students; Taking the characteristic data of each student sample as input and the corresponding sample label as output, a preset network model is used for training to obtain a student type prediction model; Obtain actual characteristic data of several students; Input the actual feature data into the student type prediction model, and output the student type to which each student belongs; Based on the student type to which the current student belongs, a matching first tutoring strategy is generated and sent to the parent end; and / or a matching second tutoring strategy is generated and sent to the student end.
64. The method for processing an electronic job according to at least one of claims 40 to 62, characterized in that: The processing method also includes: Get the syllabus content constructed in the preset authority end; The syllabus content corresponds to the assessment scope of the knowledge points; When selecting a question from the electronic question bank folder, determining whether the current question exceeds the examination scope according to the knowledge points of the current question; wherein the target question is within the examination scope; and / or, The processing method also includes: Determine the operational authority of executive entities at different levels in the same school; For the same school, the authority of the executive body to view statistical data is positively correlated with the level of the hierarchical structure; For different schools, the authority to view the statistical data range of each school is lower than that of the higher-level management department; and / or, The processing method also includes: Obtain statistical data under various statistical dimensions associated with electronic operations; Performing analysis based on the statistical data to obtain corresponding analysis results; and / or, In the preset identity recognition phase, the user's actual operation data and the actual verification data entered are obtained; When the actual verification data is the same as the preset verification data, the actual operation data is determined to be valid; otherwise, generating a third reminder message; Wherein, the actual verification data includes at least one of biometric data, preset fixed password data, and dynamic password data; The biometric data includes at least one of facial features, fingerprint features, and eye pupil features.
65. A system for obtaining an answer area, characterized in that: The acquisition system is applied in electronic operations, and the acquisition system includes: A topic determination module, used to determine the topic to be processed; An attribute information acquisition module, used to acquire the topic attribute information of the topic to be processed; A first judgment module, used for judging whether the to-be-processed question belongs to the handwriting question type based on the question attribute information, and if so, calling the region generation module; The area generation module is used to generate a handwritten answer area matching the question to be processed in the first preset blank canvas; The first preset blank canvas is a carrier for handwriting input content.
66. The system for obtaining the answer area according to claim 65, characterized in that: The handwriting answer area is used to display the corresponding stacked contents in all operation layers that have been completed and newly generated and are stacked in sequence; Among them, all the operation layers stacked in sequence are generated in sequence when different users enter the handwriting answering area in different preset use stages.
67. The system for obtaining the answer area according to claim 66, characterized in that: The operation layers generated in different preset use stages include an initial layer for the student end, a correction layer for the teacher end, a revision layer for the student end, or a revision layer for the teacher end. Wherein, the first preset blank canvas is the initial layer generated based on a canvas in a picture format; After N correction operations, the correction layer for the student end and the correction layer for the teacher end are generated again in the handwritten answer area in sequence, and N is a positive integer.
68. The system for obtaining the answer area according to claim 66, characterized in that: The handwriting answering area supports obtaining writing track related data; Wherein, the writing track associated data is stored in a preset storage space; The writing trajectory associated data includes a plurality of writing trajectory data segments, and first time frame data and / or answer writing speed data corresponding to each writing trajectory data segment; In the student terminal, the writing track data is the answer track data; In the teacher's end, the writing track data is the answer correction data.
69. The system for obtaining the answer area according to claim 68, characterized in that: The answer box area supports dynamic following display of the writing trajectory associated data in the target end.
70. The system for obtaining the answer area according to claim 68, characterized in that: The handwriting answering area supports displaying the operation layer of the student terminal in the first preset user terminal; And / or, the handwritten answer area supports replaying the answer track data stored in the preset storage space in a second preset user terminal; And / or, the handwriting answer area supports displaying the operation layer of the teacher terminal in a third preset user terminal; And / or, the handwritten answer area supports playback of the answer correction data stored in the preset storage space in a fourth preset user terminal; Wherein, the corresponding data playback mode is determined based on different selection instructions.
71. The system for obtaining the answer area according to at least one of claims 65 to 70, characterized in that: The handwriting answering area is provided with a corresponding basic function component, and the basic function component includes a plurality of basic function plug-ins.
72. The system for obtaining the answer area according to claim 71, characterized in that: The basic function plug-in includes a draft paper calling plug-in, and the draft paper calling plug-in is used to call or hide the draft paper; The draft paper area where the draft paper is located is generated following the corresponding handwritten answer area, and each handwritten answer area corresponds to one draft paper; or, all the handwritten answer areas under each of the pending questions correspond to one draft paper; The draft paper is dynamically stored in a preset storage space of the intelligent terminal.
73. The system for obtaining the answer area according to claim 72, characterized in that: The draft paper area supports obtaining draft trajectory associated data; Wherein, the draft trajectory associated data is stored in a preset storage space; The draft trajectory associated data includes a plurality of draft trajectory data segments, and second time frame data and / or draft writing speed data corresponding to each draft trajectory data segment; In the student terminal, the draft track associated data is the draft writing track data; In the teacher's end, the draft track associated data is the draft correction track data.
74. The system for obtaining the answer area according to claim 72, characterized in that: The answer box area supports dynamic following display of the draft track associated data of the draft paper in the target end.
75. The system for obtaining the answer area according to claim 72, characterized in that: The draft paper area supports displaying the draft writing track data of the student terminal in the first preset user terminal; And / or, the draft paper area supports playback of the draft writing trajectory data stored in the preset storage space in a second preset user terminal; And / or, the draft paper area supports displaying the draft correction track data of the teacher end in the third preset user end; And / or, the draft paper area supports playback of the draft correction track data stored in the preset storage space in a fourth preset user terminal; Wherein, the corresponding data playback mode is determined based on different selection instructions.
76. A system for processing electronic operations, characterized in that: The processing system is implemented according to the system for obtaining the answer area according to any one of claims 65-75, and the processing system comprises: A topic selection module is used to select a number of topics from a preset topic library based on preset learning objectives; An answer area determination module is used to generate a handwritten answer area corresponding to each question of the handwritten question type using the acquisition system; The answer position generation module is used to generate set answer positions corresponding to other remaining questions belonging to non-handwritten question types; The electronic homework acquisition module is used to acquire the electronic homework based on the question and the corresponding handwritten answer area and the set answer position.
77. The electronic operation processing system as claimed in claim 76, characterized in that: The processing system further comprises: The answering progress data acquisition module is used to obtain the answering progress data corresponding to the question and send it to a preset storage space; The answering progress data includes first characterization data of whether the student terminal has entered the question, and answering track data of the student terminal in the answering area of the question and / or draft writing track data of the draft paper; A first display control module, used for controlling the shared display of a first current display interface of the student terminal in a first preset user terminal when the first representation data represents that the student terminal enters the question; A first playback control module, configured to control the playback of the answer track data of the student terminal and / or the draft writing track data of the draft paper in the first preset user terminal after the student terminal submits the question; or, A correction progress data acquisition module, used to acquire the correction progress data corresponding to the question and send it to the preset storage space; The correction progress data includes second characterization data of whether the teacher terminal enters the question, and correction track data of the teacher terminal on the answer area of the question and / or correction track data of the draft on the draft paper; A second display control module, used for controlling the second current display interface of the teacher terminal to be displayed on a second preset user terminal when the second representation data represents that the teacher terminal has entered the question; The second playback control module is used to control the playback of the teacher's terminal's correction track data and / or the draft correction track data of the draft paper in the second preset user terminal after the teacher's terminal completes the question correction.
78. The electronic operation processing system as claimed in claim 76, characterized in that: The processing system further comprises: The third display control module is configured to perform at least one of the following operations for the plurality of electronic jobs in the same batch: Obtaining the completion progress of each electronic homework, dynamically tracking and displaying each electronic homework, obtaining each electronic homework for static display, and obtaining each electronic homework for playback; and / or, The third playback control module is used to control the static display or playback of different handwritten answer contents corresponding to the same handwritten question in different electronic assignments.
79. An electronic operation operation and management platform, characterized in that: The electronic operation operation and management platform includes the electronic operation processing system according to any one of claims 76-78, and a plurality of preset user terminals for communication interaction; The plurality of preset user terminals include a student homework application terminal in a student device, and a teacher homework application terminal in a teacher device.
80. The electronic operation operation and management platform as claimed in claim 79, characterized in that: The plurality of preset user terminals also include at least one of the following application terminals: A parent application terminal in a parent device, a first education supervision application terminal in a school education supervision device, and a second education supervision application terminal in an education bureau education supervision device; The parent application terminal communicates and interacts with the student homework application terminal and the teacher homework application terminal respectively; The first education supervision application end and the second education supervision application end communicate and interact with the student homework application end and the teacher application end respectively.
81. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the method for obtaining the answer area according to any one of claims 1-21; or, it implements the method for processing electronic homework according to any one of claims 22-64.
82. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the method for obtaining the answer area according to any one of claims 1-21; or, it implements the method for processing electronic homework according to any one of claims 22-64.
Citation Information
Cited By
User identity authentication method based on LLM model, server, client and program product
CN119227034A
Intelligent writing and correcting method and system
CN120670789A
Digital marking method and device for paper homework
CN121074933A
