Card punching management method, device and equipment for offline operation of online education training camp

Abstract

The invention discloses a card punching management method, device and equipment for offline operation of an online education training camp, and relates to the technical field of artificial intelligence. The method comprises the following steps: firstly, extracting answer content and answer duration of each test question from offline homework of a target student, obtaining text similarity between the answer content and a standard answer for each test question, and segmenting the test question according to a corresponding class-following monitoring video; obtaining the confidence coefficient that the class attending state of the target student on the corresponding knowledge point is a careful class attending state, then importing the answering duration, the text similarity and the confidence coefficient corresponding to each test question into a homework plagiarism classification model, and outputting another confidence coefficient that the offline homework of the target student has plagiarism behavior. And finally, according to a comparison result of the other confidence coefficient and a preset threshold value, determining whether the target student has an offline homework plagiarism condition, if so, pushing a new offline homework to the target student again, and otherwise, receiving the offline homework and completing card punching.

Description

Technical Field

[0001] The present invention belongs to the technical field of artificial intelligence, and particularly relates to an offline homework check-in management method, device and equipment for an online education training camp. Background Art

[0002] As is well known, the learning effect of a group is much greater than that of an individual. The online education training camp activities launched by training institutions are a new online hybrid training model of "online knowledge input + community service". With the online community as the carrier, gathering students with the same learning needs can not only efficiently and systematically complete learning tasks, but also create more opportunities for communication and sharing, greatly enhancing the learning experience.

[0003] Currently, after the training camp courses, corresponding offline homework is generally assigned, and students are required to submit the completed offline homework within the specified check-in deadline. However, due to the lack of supervision of students' offline homework, it is impossible to distinguish whether there is plagiarism in the students' offline homework, which will affect the accuracy of the training institution's grasp of the students' learning situation. Summary of the Invention

[0004] The purpose of the present invention is to provide an offline homework check-in management method, device, computer device, computer-readable storage medium and computer program product for an online education training camp, so as to solve the problem that the accuracy of the training institution's grasp of the students' learning situation is limited because it is impossible to distinguish whether there is plagiarism in the students' offline homework.

[0005] To achieve the above purpose, the present invention adopts the following technical solutions:

[0006] In the first aspect, an offline homework check-in management method for an online education training camp is provided, including:

[0007] After receiving the first offline homework of an online education training camp submitted by a target student, extracting the first answer content and answer duration of each test question from the first offline homework of the online education training camp;

[0008] For each test question, calculating a first text similarity between the corresponding standard answer content and the corresponding first answer content according to the corresponding standard answer content;

[0009] For each test question, intercepting a segment of the in-class listening monitoring video with the acquisition period consistent with the display period of the lecture page of the corresponding knowledge point from the in-class listening monitoring video of the target student according to the corresponding knowledge point, and importing the segment of the in-class listening monitoring video into a pre-trained listening state classification model based on the first machine learning algorithm, and outputting a first confidence level that the listening state of the target student on the corresponding knowledge point is a serious listening state;

[0010] Importing the answering time corresponding to each test question, the first text similarity, and the first confidence into a pre-trained homework plagiarism classification model based on a second machine learning algorithm, and outputting a second confidence that the offline homework of the first online education training camp contains plagiarism;

[0011] Determining whether the second confidence level exceeds a preset first confidence level threshold;

[0012] If so, corresponding new test questions are re-extracted for the knowledge points to which the test questions belong, and all the new test questions are pushed to the target students as new offline homework for the target students; otherwise, the submission timestamp of the offline homework of the first online education training camp is used as the offline homework check-in timestamp of the target students.

[0013] Based on the above invention content, a new solution for managing students' offline homework punch-in based on in-class lecture monitoring video and answer data is provided, that is, first, the answer content and answer time of each test question are extracted from the offline homework of the target student, and for each test question, the text similarity between the answer content and the standard answer is obtained, and the in-class lecture monitoring video segmentation is consistent with the display time period of the lecture page corresponding to the corresponding knowledge point, and the confidence that the target student's listening status at the corresponding knowledge point is a serious listening state is obtained, and then the answer time, text similarity and confidence corresponding to each test question are imported into the homework plagiarism classification model, and another confidence that the target student's offline homework has plagiarism is output, and finally, according to the comparison result of the other confidence and the preset threshold, it is determined whether the target student has plagiarized the offline homework, and if so, the new offline homework is re-pushed to the target student, otherwise the offline homework is received and completed and punched in, so that it can automatically distinguish whether the student's offline homework has plagiarism, improve the accuracy of the teaching and training institutions in grasping the students' learning situation, and facilitate practical application and promotion.

[0014] In a possible design, the submission timestamp of the offline homework of the first online education training camp is used as the offline homework check-in timestamp of the target student, including:

[0015] For each student who checked in previously, determine whether the target student has plagiarized the offline homework of the corresponding student;

[0016] If it is determined that the target student has plagiarized the offline homework of any prior clock-in student, then for the knowledge points to which the respective test questions belong, new corresponding test questions are re-extracted, and all the new test questions are pushed to the target student as the new offline homework of the target student; otherwise, the submission timestamp of the offline homework of the first online education training camp is used as the offline homework clock-in timestamp of the target student.

[0017] In a possible design, for each prior clock-in student, determining whether there is a situation where the target student has plagiarized the offline homework of the corresponding student includes:

[0018] Extracting the second answer content of each test question from the offline homework of the second online education training camp submitted by a certain prior clock-in student;

[0019] For each test question, calculating the second text similarity between the second answer content and the corresponding first answer content according to the corresponding second answer content;

[0020] Importing the answer duration, the second text similarity, and the first confidence level corresponding to each test question into the homework plagiarism classification model, and outputting the third confidence level that there is a plagiarism behavior in the offline homework of the first online education training camp;

[0021] Judging whether the third confidence level exceeds a preset second confidence level threshold;

[0022] If so, it is determined that there is a situation where the target student has plagiarized the offline homework of the certain prior clock-in student; otherwise, it is determined that there is no situation where the target student has plagiarized the offline homework of the certain prior clock-in student.

[0023] In a possible design, for each test question, calculating the first text similarity between the standard answer content and the corresponding first answer content according to the corresponding standard answer content includes:

[0024] Performing word segmentation on the first answer content of a certain test question to obtain a first word set, and performing the word segmentation on the standard answer content of the certain test question to obtain a second word set;

[0025] Calculating the first text similarity T between the standard answer content of the certain test question and the first answer content of the certain test question according to the following formula sim,1 :

[0026] T sim,1 = max(T TF-IDF,sim × T JSC,sim , T MED,sim × T JSC,sim )

[0027] In the formula, T TF-IDF,sim represents the similarity degree between the first word set and the second word set calculated based on the term frequency-inverse document frequency TF-IDF, and T MED,sim represents the difference degree between the first word set and the second word set calculated based on the edit distance MED, and T JSC,sim represents the similarity degree between the first word set and the second word set calculated based on the Jaccard similarity coefficient, and max() represents the maximum value function.

[0028] In a possible design, the first machine learning algorithm adopts an artificial intelligence algorithm based on a support vector machine, a decision tree, or a random forest, and the second machine learning algorithm adopts an artificial intelligence algorithm based on the K-nearest neighbor method, the stochastic gradient descent method, a multi-layer perceptron, a backpropagation neural network, or a radial basis function network.

[0029] In a possible design, for the knowledge points to which the respective test questions belong, corresponding new test questions are redrawn, including:

[0030] For the knowledge points to which the respective test questions belong, according to the first confidence level that the listening state of the target student for the corresponding knowledge point is a serious listening state, the corresponding question-setting difficulty negatively correlated with the first confidence level is determined, and corresponding new test questions are drawn from the corresponding question bank according to the question-setting difficulty.

[0031] In a second aspect, an offline homework punching management device for an online education training camp is provided, including an answer data extraction unit, a text similarity calculation unit, a listening state classification unit, a homework plagiarism classification unit, a judgment unit, and a judgment response unit;

[0032] The answer data extraction unit is configured to, after receiving the first offline homework of the online education training camp submitted by the target student, extract the first answer content and answer duration of each test question from the first offline homework of the online education training camp;

[0033] The text similarity calculation unit is communicatively connected to the answer data extraction unit, and is configured to, for each test question, calculate a first text similarity between the corresponding standard answer content and the corresponding first answer content according to the corresponding standard answer content;

[0034] The class-attending status classification unit is communicatively connected to the answer data extraction unit. For each test question, according to the corresponding knowledge point it belongs to, a segment of the in-class monitoring video of the target student is intercepted from the in-class monitoring video of the target student, where the acquisition period of the segment is consistent with the display period of the lecture page corresponding to the knowledge point, and this segment of the in-class monitoring video is imported into a pre-trained class-attending status classification model based on the first machine learning algorithm, and the first confidence level that the class-attending status of the target student for the corresponding knowledge point belongs to the serious class-attending status is output;

[0035] The homework plagiarism classification unit is communicatively connected to the answer data extraction unit, the text similarity calculation unit, and the class-attending status classification unit respectively, and is used to import the answer duration, the first text similarity, and the first confidence level corresponding to each test question into a pre-trained homework plagiarism classification model based on the second machine learning algorithm, and output the second confidence level that there is a plagiarism behavior in the offline homework of the first online education training camp;

[0036] The judgment unit is communicatively connected to the homework plagiarism classification unit, and is used to judge whether the second confidence level exceeds a preset first confidence threshold;

[0037] The determination response unit is communicatively connected to the judgment unit, and is used to, when it is determined that the second confidence level exceeds the first confidence threshold, re-extract corresponding new test questions for the knowledge points to which each test question belongs, and push all the new test questions as the new offline homework of the target student to the target student, and when it is determined that the second confidence level does not exceed the first confidence threshold, use the submission timestamp of the offline homework of the first online education training camp as the offline homework check-in timestamp of the target student.

[0038] In a third aspect, the present invention provides a computer device, including a memory, a processor, and a transceiver that are communicatively connected in sequence. Among them, the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the offline homework check-in management method of the online education training camp as described in the first aspect or any possible design in the first aspect.

[0039] In a fourth aspect, the present invention provides a computer-readable storage medium, on which instructions are stored. When the instructions are run on a computer, the offline homework check-in management method of the online education training camp as described in the first aspect or any possible design in the first aspect is executed.

[0040] Fifth aspect, the present invention provides a computer program product, including a computer program or instructions, which, when executed by a computer, implement the online education training camp offline homework check-in management method as described in the first aspect or any possible design in the first aspect.

[0041] Beneficial effects of the above solution:

[0042] (1) The present invention creatively provides a new solution for managing students' offline homework check-in based on in-class lecture monitoring videos and answer data. That is, first extract the answer content and answer duration of each test question from the offline homework of the target student, and for each test question, obtain the text similarity between the answer content and the standard answer, and according to the in-class lecture monitoring video segments where the collection time period is consistent with the display time period of the corresponding lecture page of the corresponding knowledge point, obtain the confidence level that the target student is in a serious listening state for the corresponding knowledge point. Then import the answer duration, text similarity, and confidence level corresponding to each test question into the homework plagiarism classification model, output another confidence level that the target student's offline homework has a plagiarism behavior. Finally, according to the comparison result between this another confidence level and the preset threshold, determine whether the target student has an offline homework plagiarism situation. If so, push a new offline homework to the target student again, otherwise receive the offline homework to complete the check-in. In this way, it can automatically distinguish whether there is a plagiarism situation in the offline homework of students, improve the accuracy of the training institution's grasp of students' learning situations, and facilitate practical application and promotion. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0044] Figure 1 It is a flowchart of the online education training camp offline homework check-in management method provided by the embodiment of the present application.

[0045] Figure 2 It is a structural diagram of the online education training camp offline homework check-in management device provided by the embodiment of the present application.

[0046] Figure 3 It is a structural diagram of the computer device provided by the embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For those of ordinary skill in the art, other embodiments can be obtained based on these embodiments without creative efforts. It should be noted here that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation on the present invention.

[0048] It should be understood that although terms such as first and second may be used herein to describe various objects, these objects should not be limited by these terms. These terms are only used to distinguish one object from another. For example, the first object may be referred to as the second object, and similarly, the second object may be referred to as the first object, without departing from the scope of the exemplary embodiments of the present invention.

[0049] It should be understood that for the term "and / or" that may appear in this document, it is only a description of the association relationship between associated objects, indicating that three relationships may exist. For example, A and / or B may mean: A exists alone, B exists alone, or A and B exist simultaneously; for another example, A, B and / or C may mean that any one of A, B and C exists or any combination of them exists; for the term " / and" that may appear in this document, it is a description of another association object relationship, indicating that two relationships may exist. For example, A / and B may mean: A exists alone or A and B exist simultaneously; in addition, for the character " / " that may appear in this document, generally it means that the associated objects before and after are in an "or" relationship.

[0050] Embodiment:

[0051] As Figure 1 shown, the offline homework clock-in management method of the online education training camp provided in the first aspect of this embodiment can, but is not limited to, be executed by a computer device with certain computing resources, such as a platform server, a personal computer (Personal Computer, PC, referring to a multi-purpose computer suitable for personal use in terms of size, price and performance; desktop computers, laptops to small laptops, tablets and ultrabooks, etc. all belong to personal computers), a smart phone, a personal digital assistant (Personal Digital Assistant, PDA) or a wearable device and other electronic devices. As Figure 1 shown, the offline homework clock-in management method of the online education training camp can, but is not limited to, include the following steps S1 to S6.

[0052] S1. After receiving the offline homework of the first online education training camp submitted by the target trainee, extract the first answer content and answer duration of each test question from the offline homework of the first online education training camp.

[0053] In step S1, the target trainee is the object to be managed; the offline homework of the first online education training camp is the answer result of the offline homework assigned after the trainee completes the training camp courses, and the operation details of existing examination software can be specifically referred to for homework distribution, answering, and submission. The first answer content is the answer result data of the target trainee for the corresponding test question, which can be regularly extracted based on the preset answer position. For example, extract the multiple-choice options, fill-in-the-blank input characters, or true / false options from the originally empty brackets; the answer duration is the answer process data of the target trainee for the corresponding test question, and the display duration of the test question page recorded by the examination software can be used as the answer duration (existing examination software will automatically or manually jump to the answer display page of the next test question after finishing a test question, so the display duration of the test question page can be used as the answer duration).

[0054] S2. For each test question, calculate the first text similarity between the corresponding standard answer content and the first answer content.

[0055] In step S2, the standard answer content of each test question is generated synchronously when the corresponding test question is generated, so it can be regularly read. Preferably, for each test question, calculating the first text similarity between the corresponding standard answer content and the first answer content includes, but is not limited to, the following steps S21 - S22.

[0056] S21. Perform word segmentation on the first answer content of a certain test question to obtain a first word set, and perform the word segmentation on the standard answer content of the certain test question to obtain a second word set.

[0057] In step S21, the word segmentation can specifically but is not limited to using the jieba word segmentation tool (it is a widely used Chinese word segmentation tool that supports multiple word segmentation modes, including the precise mode, full mode, and search engine mode. Among them, the precise mode aims to cut the sentence most precisely and is suitable for text analysis; the full mode scans all possible word-forming units in the sentence, with a faster speed but does not solve ambiguity; the search engine mode performs secondary segmentation on long words based on the precise mode to improve the recall rate and is especially suitable for search engine applications) to achieve.

[0058] S22. Calculate the first text similarity T between the standard answer content of a certain test question and the first answer content of the certain test question according to the following formula sim,1 :

[0059] T sim,1 = max(T TF-IDF,sim × T JSC,sim , T MED,sim × T JSC,sim )

[0060] In the formula, T TF-IDF,sim represents the similarity degree between the first word set and the second word set calculated based on the term frequency-inverse document frequency TF-IDF (Term Frequency–Inverse Document Frequency, which is a statistical method used to evaluate the importance of a word for a document set or a single document in a corpus), T MED,sim represents the difference degree between the first word set and the second word set calculated based on the minimum edit distance MED (Minimum Edit Distance, which is a quantitative measurement of the difference degree between two strings. The measurement method is to see how many times of processing are required at least to change one string into another string), T JSC,sim represents the similarity degree between the first word set and the second word set calculated based on the Jaccard Similarity Coefficient (which is used to measure the similarity between two sets and is defined as the number of elements in the intersection of two sets divided by the number of elements in the union of two sets), and max() represents the maximum value function.

[0061] In the step S22, the specific calculation formulas of the above intermediate parameters T TF-IDF,sim , T MED,sim and T JSC,sim are all existing technical means. Although T TF-IDF,sim will be a cosine value in the dimension of term frequency-inverse document frequency, T MED,sim in the dimension of minimum edit distance and T JSC,sim in the dimension of Jaccard similarity coefficient will be normalized respectively, so that their values are all in [0, 1]. However, it should be noted that the values given in each dimension describe different degrees of similarity / difference. Even if the same values are given in each dimension, their similarity / difference degrees are different. How to aggregate these three intermediate parameters in different situations is an important issue. Consider T TF-IDF,simThe representational meaning, and its value is affected by word frequency. That is, when the content of the standard answer or the first response is short, the word frequency of each word may be 1. At this time, it is not possible to well distinguish the importance of each word. Therefore, T TF-IDF,sim will be not conducive to calculating the similarity between too short texts (i.e., between the content of the standard answer and the first response). And T MED,sim is exactly the opposite compared with T TF-IDF,sim . It is more suitable for evaluating the difference between shorter texts because there are limited words cut from shorter texts, and it only takes fewer calculation steps to find the mapping of the second word (i.e., the word in the second word set) in the first word set. And T JSC,sim will be an index insensitive to text length, that is, regardless of the length of the text, it always determines the set difference by the size of the intersection. Therefore, through the above formula, when the text is short, T MED,sim ×T JSC,sim (generally, T MED,sim is greater than T TF-IDF,sim at this time) can be used to comprehensively measure the similarity / difference between the first word set and the second word set. And when the text is long, T TF-IDF,sim ×T JSC,sim (generally, T TF-IDF,sim is greater than T MED,sim at this time) can be used to comprehensively measure the similarity / difference between the first word set and the second word set. That is, regardless of the text length, the accuracy of index aggregation (i.e., the first text similarity T sim,1 ) can be ensured.

[0062] S3. For each of the test questions, intercept a segment of the classroom listening monitoring video of the target student from the classroom listening monitoring video of the target student, where the acquisition period is consistent with the display period of the corresponding lecture page of the knowledge point according to the corresponding knowledge point to which it belongs, and import the segment of the classroom listening monitoring video into the listening state classification model based on the first machine learning algorithm and already pre-trained, and output the first confidence level that the listening state of the target student on the corresponding knowledge point belongs to the serious listening state.

[0063] In the step S3, the corresponding relationships between the respective test questions and the knowledge points can be established automatically or manually when the corresponding test questions are generated. For example, for a certain knowledge point, multiple test questions corresponding to it and covering different question types are automatically generated by using AI (Artificial Intelligence) technology. The in-class listening monitoring video is a monitoring video regularly recorded for the listening performance of the target student in the online classroom corresponding to the offline homework of the first online education training camp (which can be regularly recorded by the camera on the student side and uploaded to the local). For example, if the display period of the lecture page corresponding to a certain knowledge point in the online classroom is from 10:00 to 10:05, then for the certain knowledge point, it is necessary to extract a segment of the in-class listening monitoring video with a time period between 10:00 and 10:05 from the in-class listening monitoring video of the target student. Since the segment of the in-class listening monitoring video will present the micro-expression features of the target student during the display period of the lecture page: behavior features such as concentration or distraction, etc., therefore, based on a certain amount of samples of the in-class listening monitoring video segments and sample labels (such as the state of listening carefully or not listening carefully), through the conventional calibration and verification modeling method (the specific process includes the calibration process and the verification process of the model, that is, first comparing the model simulation results with the measured data, and then adjusting the model parameters according to the comparison results to make the simulation results coincide with the actual situation), the listening state classification model can be pre-trained. In addition, specifically, the first machine learning algorithm adopts an artificial intelligence algorithm based on support vector machine, decision tree or random forest, etc. (which is a core artificial intelligence algorithm that specifically studies how a computer simulates or realizes human learning behavior to acquire new knowledge or skills, reorganize the existing knowledge structure to continuously improve its own performance, and is the fundamental way to make a computer intelligent).

[0064] S4. Import the answering duration, the first text similarity, and the first confidence corresponding to the respective test questions into the pre-trained homework plagiarism classification model based on the second machine learning algorithm, and output the second confidence that there is a plagiarism behavior in the offline homework of the first online education training camp.

[0065] In the step S4, since the answering duration and the first text similarity can reflect the first mastery level of the target student on the knowledge points corresponding to the test questions from the perspective of answering questions. For example, the faster and more accurate the target student answers the questions, the higher the mastery level, and vice versa. The first confidence level can reflect the second mastery level of the target student on the knowledge points corresponding to the test questions from the perspective of attending classes. For example, the more seriously the target student attends classes on a certain knowledge point (at this time, the first confidence level is higher), it can reflect that the target student has a higher mastery level of this certain knowledge point, and vice versa. Therefore, based on the matching result of the first mastery level and the second mastery level, it can be identified whether there is a plagiarism behavior through big data analysis: if they match, there is no plagiarism behavior; if they do not match, there may be a plagiarism behavior. Thus, after pre-training the homework plagiarism classification model through a conventional calibration and verification modeling method based on a certain amount of sample data (i.e., sample answering duration, sample similarity, sample confidence level, etc.) and label data (such as plagiarism or non-plagiarism), the homework plagiarism classification model can be applied to estimate the second confidence level that there is a plagiarism behavior in the offline homework of the first online education training camp. In addition, specifically, the second machine learning algorithm uses an artificial intelligence algorithm based on the K-nearest neighbor method, stochastic gradient descent method, multi-layer perceptron, backpropagation neural network or radial basis function network.

[0066] S5. Determine whether the second confidence level exceeds a preset first confidence level threshold.

[0067] In the step S5, the first confidence level threshold can be exemplified as 62%.

[0068] S6. If so, for the knowledge points corresponding to each test question, re-extract the corresponding new test questions, and push all the new test questions to the target student as the new offline homework of the target student, otherwise use the submission timestamp of the offline homework of the first online education training camp as the offline homework check-in timestamp of the target student.

[0069] In step S6, if the second confidence level exceeds the first confidence threshold, it can be considered that there is a plagiarism behavior in the offline homework of the first online education training camp submitted by the target student, and it is necessary to re-draw questions for supplementary completion to ensure the target student's mastery of the knowledge points of each test question, which is conducive to their ability to keep up with the subsequent online education training camp teaching courses and improve the activity effect of the online education training camp. Preferably, for the knowledge points of each test question, corresponding new test questions are re-drawn, including but not limited to: for the knowledge points of each test question, according to the first confidence level that the target student's listening state for the corresponding knowledge point is a serious listening state, the corresponding question-setting difficulty negatively correlated with this first confidence level is determined, and corresponding new test questions are drawn from the corresponding question bank according to this question-setting difficulty. The purpose of establishing the negative correlation between the first confidence level and the question-setting difficulty is as follows: if the target student listens more attentively to a certain knowledge point (i.e., the first confidence level is higher), it can be reflected that the target student has a higher mastery of this certain knowledge point. Therefore, easier test questions can be pushed for offline routine training; while if the target student listens less attentively to a certain knowledge point (i.e., the first confidence level is lower), it can be reflected that the target student has a lower mastery of this certain knowledge point. Therefore, more difficult test questions can be pushed for offline focused training to improve the target student's mastery of this certain knowledge point. Furthermore, through this targeted question-setting method, the mastery of knowledge points by all students can be balanced, which is further conducive to their ability to keep up with the subsequent online education training camp teaching courses.

[0070] In step S6, in order to further distinguish whether there is mutual plagiarism in the offline homework among students, preferably, the submission timestamp of the first online education training camp offline homework is used as the offline homework check-in timestamp of the target student, including but not limited to the following steps S61 to S62.

[0071] S61. For each prior check-in student, determine whether there is a situation where the target student plagiarizes the offline homework of the corresponding student.

[0072] In step S61, the prior check-in student is another target student who has completed the identification of homework plagiarism based on the foregoing steps S1 to S6 and successfully completed the check-in (i.e., determined the offline homework check-in timestamp) because it is determined that there is no homework plagiarism behavior. Specifically, for each prior check-in student, determine whether there is a situation where the target student plagiarizes the offline homework of the corresponding student, including but not limited to the following steps S611 to S615.

[0073] S611. Extract the second answer content of each test question from the second online education training camp offline homework submitted by a certain prior check-in student.

[0074] In step S611, the second online education training camp offline homework is the answer result of the offline homework assigned after completing the training camp course by the student who previously checked in, which can also refer to the operation details of the existing examination software to implement homework distribution, answering and submission. The second answer content is the answer result data of the corresponding test question by the student who previously checked in, which can also be conventionally extracted based on the preset answer position, such as extracting multiple-choice question options, fill-in-the-blank question input characters or true-or-false question options from brackets that were originally empty brackets.

[0075] S612. For each of the test questions, calculate the second text similarity between the second answer content and the corresponding first answer content according to the corresponding second answer content.

[0076] S613. Import the answering time corresponding to each test question, the second text similarity and the first confidence into the homework plagiarism classification model, and output a third confidence that there is plagiarism in the offline homework of the first online education training camp.

[0077] S614. Determine whether the third confidence level exceeds a preset second confidence level threshold.

[0078] In the step S614, the second confidence threshold can also be 62%. In addition, the specific technical details of the above steps S612 to S614 can be conventionally derived with reference to the above steps S2 and S4 to S5, and will not be repeated here.

[0079] S615. If so, it is determined that the target student plagiarizes the offline homework of the student who checked in earlier; otherwise, it is determined that the target student does not plagiarize the offline homework of the student who checked in earlier.

[0080] S62. If it is determined that the target student has plagiarized the offline homework of any student who has checked in previously, then new test questions corresponding to the knowledge points of each test question are re-extracted, and all the new test questions are pushed to the target student as the new offline homework of the target student. Otherwise, the submission timestamp of the offline homework of the first online education training camp is used as the check-in timestamp of the target student's offline homework.

[0081] Based on the online education training camp offline homework check-in management method described in the foregoing steps S1 to S6, a new solution for managing the offline homework check-in of students based on the in-class lecture monitoring video and answering data is provided. That is, first, the answering content and answering duration of each test question are extracted from the offline homework of the target student, and for each test question, the text similarity between the answering content and the standard answer is obtained, and according to the in-class lecture monitoring video segments where the collection period is consistent with the display period of the corresponding lecture page of the corresponding knowledge point, the confidence level that the target student is in a serious listening state for the corresponding knowledge point is obtained. Then, the answering duration, text similarity, and confidence level corresponding to each test question are imported into the homework plagiarism classification model, and another confidence level that the offline homework of the target student has a plagiarism behavior is output. Finally, according to the comparison result between this another confidence level and the preset threshold, it is determined whether the target student has a situation of plagiarizing the offline homework. If so, a new offline homework is pushed to the target student again, otherwise, the offline homework completion check-in is received. In this way, it can automatically distinguish whether there is a plagiarism situation in the offline homework of students, improve the accuracy of the training institution's grasp of the learning situation of students, and facilitate practical application and promotion.

[0082] As Figure 2 shown, in the second aspect of this embodiment, a virtual device for implementing the online education training camp offline homework check-in management method described in the first aspect is provided, including an answering data extraction unit, a text similarity calculation unit, a listening state classification unit, a homework plagiarism classification unit, a judgment unit, and a determination response unit;

[0083] The answering data extraction unit is used to extract the first answering content and answering duration of each test question from the first online education training camp offline homework after receiving the first online education training camp offline homework submitted by the target student;

[0084] The text similarity calculation unit is communicatively connected to the answering data extraction unit, and is used to calculate the first text similarity between the standard answer content and the corresponding first answering content for each test question according to the corresponding standard answer content;

[0085] The listening state classification unit is communicatively connected to the answering data extraction unit. For each test question, according to the corresponding knowledge point, the in-class lecture monitoring video segments where the collection period is consistent with the display period of the corresponding lecture page of the knowledge point are intercepted from the in-class lecture monitoring video of the target student, and the in-class lecture monitoring video segments are imported into the listening state classification model that is based on the first machine learning algorithm and has been pre-trained, and the first confidence level that the target student is in a serious listening state for the corresponding knowledge point is output;

[0086] The homework plagiarism classification unit is communicatively connected to the answer data extraction unit, the text similarity calculation unit, and the class attendance status classification unit respectively, and is used to import the answer duration, the first text similarity, and the first confidence level corresponding to each test question into a pre-trained homework plagiarism classification model based on a second machine learning algorithm, and output a second confidence level that there is a plagiarism behavior in the offline homework of the first online education training camp;

[0087] The judgment unit is communicatively connected to the homework plagiarism classification unit, and is used to judge whether the second confidence level exceeds a preset first confidence level threshold;

[0088] The determination response unit is communicatively connected to the judgment unit, and is used to, when it is determined that the second confidence level exceeds the first confidence level threshold, re-extract corresponding new test questions for the knowledge points to which each test question belongs, and push all the new test questions as the new offline homework of the target student to the target student, and when it is determined that the second confidence level does not exceed the first confidence level threshold, use the submission timestamp of the offline homework of the first online education training camp as the offline homework check-in timestamp of the target student.

[0089] For the working process, working details and technical effects of the foregoing device provided in the second aspect of this embodiment, reference may be made to the online education training camp offline homework check-in management method described in the first aspect, and details are not described herein again.

[0090] As Figure 3 shown, a computer device for executing the online education training camp offline homework check-in management method described in the first aspect is provided in the third aspect of this embodiment, including a memory, a processor, and a transceiver that are communicatively connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the online education training camp offline homework check-in management method described in the first aspect. Specifically, for example, the memory may include, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a flash memory, a first-in first-out memory (FIFO), and / or a first-in last-out memory (FILO), etc.; the processor may include, but is not limited to, a microprocessor of the STM32F105 series. In addition, the computer device may further include, but is not limited to, a power module, a display screen, and other necessary components.

[0091] For the working process, working details and technical effects of the foregoing computer device provided in the third aspect of this embodiment, reference may be made to the online education training camp offline homework check-in management method described in the first aspect, which will not be elaborated herein.

[0092] In the fourth aspect of this embodiment, there is provided a computer-readable storage medium storing instructions including the online education training camp offline homework check-in management method described in the first aspect, that is, instructions are stored on the computer-readable storage medium, and when the instructions run on a computer, the online education training camp offline homework check-in management method described in the first aspect is executed. Among them, the computer-readable storage medium refers to a carrier for storing data, and may include, but is not limited to, computer-readable storage media such as floppy disks, optical discs, hard disks, flash memories, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0093] For the working process, working details and technical effects of the foregoing computer-readable storage medium provided in the fourth aspect of this embodiment, reference may be made to the online education training camp offline homework check-in management method described in the first aspect, which will not be elaborated herein.

[0094] In the fifth aspect of this embodiment, there is provided a computer program product including a computer program or instructions, and when the computer program or the instructions are executed by a computer, the online education training camp offline homework check-in management method described in the first aspect is implemented. Among them, the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0095] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An offline homework check-in management method for an online education training camp, characterized in that, Including: After receiving the offline homework of the first online education training camp submitted by the target trainee, extracting the first answer content and answer duration of each test question from the offline homework of the first online education training camp; For each of the test questions, calculating the first text similarity between the corresponding standard answer content and the corresponding first answer content according to the corresponding standard answer content; For each of the test questions, intercepting a segmented video of the in-class listening monitoring of the target trainee with the acquisition period consistent with the display period of the lecture page of the corresponding knowledge point from the in-class listening monitoring video of the target trainee according to the corresponding knowledge point to which it belongs, and importing the segmented video of the in-class listening monitoring into a pre-trained listening state classification model based on the first machine learning algorithm, and outputting the first confidence level that the listening state of the target trainee on the corresponding knowledge point to which it belongs is a serious listening state; Importing the answer duration, the first text similarity, and the first confidence level corresponding to each of the test questions into a pre-trained homework plagiarism classification model based on the second machine learning algorithm, and outputting the second confidence level that the offline homework of the first online education training camp has a plagiarism behavior; Judging whether the second confidence level exceeds a preset first confidence level threshold; If so, re-extracting corresponding new test questions for the knowledge points to which each of the test questions belongs, and pushing all the new test questions as the new offline homework of the target trainee to the target trainee, otherwise using the submission timestamp of the offline homework of the first online education training camp as the offline homework clock-in timestamp of the target trainee.

2. The offline homework check-in management method for the online education training camp according to claim 1, wherein Using the submission timestamp of the offline homework of the first online education training camp as the offline homework clock-in timestamp of the target trainee includes: For each of the previously clocked-in trainees, judging whether there is a situation where the target trainee plagiarizes the offline homework of the corresponding trainee; If it is determined that there is a situation where the target trainee plagiarizes the offline homework of any of the previously clocked-in trainees, re-extracting corresponding new test questions for the knowledge points to which each of the test questions belongs, and pushing all the new test questions as the new offline homework of the target trainee to the target trainee, otherwise using the submission timestamp of the offline homework of the first online education training camp as the offline homework clock-in timestamp of the target trainee.

3. The offline homework check-in management method for the online education training camp according to claim 2, wherein, For each of the previously clocked-in trainees, judging whether there is a situation where the target trainee plagiarizes the offline homework of the corresponding trainee includes: Extracting the second answer content of each of the test questions from the offline homework of the second online education training camp submitted by a certain previously clocked-in trainee; For each of the test questions, calculating the second text similarity between the corresponding second answer content and the corresponding first answer content according to the corresponding second answer content; Importing the answer duration, the second text similarity, and the first confidence level corresponding to each of the test questions into the homework plagiarism classification model, and outputting the third confidence level that the offline homework of the first online education training camp has a plagiarism behavior; Judging whether the third confidence level exceeds a preset second confidence level threshold; If so, it is determined that there is a situation where the target student plagiarizes the offline homework of a certain previous clock-in student; otherwise, it is determined that there is no such situation where the target student plagiarizes the offline homework of a certain previous clock-in student.

4. The offline homework check-in management method for the online education training camp according to claim 1, characterized in that For each of the test questions, calculate the first text similarity between the corresponding standard answer content and the corresponding first answer content according to the corresponding standard answer content, including: Perform word segmentation on the first answer content of a certain test question to obtain a first word set, and perform the word segmentation on the standard answer content of the certain test question to obtain a second word set; The first text similarity T between the standard answer content of a certain test question and the first answer content of the certain test question is calculated according to the following formula sim,1 :[[]]END]] T sim,1 = max(T TF-IDF,sim × T JSC,sim , T MED,sim × T JSC,sim ) Wherein, T TF-IDF,sim represents the similarity degree between the first word set and the second word set calculated based on the term frequency-inverse document frequency TF-IDF, and T MED,sim represents the difference degree between the first word set and the second word set calculated based on the edit distance MED, and T JSC,sim represents the similarity degree between the first word set and the second word set calculated based on the Jaccard similarity coefficient, and max() represents the maximum value function.

5. The offline homework punch-in management method for online education training camp according to claim 1 is characterized in that: The first machine learning algorithm uses an artificial intelligence algorithm based on support vector machine, decision tree or random forest, and the second machine learning algorithm uses an artificial intelligence algorithm based on K-nearest neighbor method, stochastic gradient descent method, multi-layer perceptron, backpropagation neural network or radial basis function network.

6. The offline homework check-in management method for the online education training camp according to claim 1, characterized in that For the knowledge points to which each of the test questions belongs, re-extract the corresponding new test questions, including: For the knowledge points to which each of the test questions belongs, determine the corresponding question difficulty negatively correlated with the first confidence level according to the first confidence level that the target student's listening state for the corresponding knowledge point is a serious listening state, and extract the corresponding new test questions from the corresponding question bank according to the question difficulty.

7. An offline homework check-in management device for an online education training camp, characterized in that, It includes an answer data extraction unit, a text similarity calculation unit, a listening state classification unit, a homework plagiarism classification unit, a judgment unit and a determination response unit; The answer data extraction unit is used to extract the first answer content and answer duration of each test question from the first offline homework of the online education training camp after receiving the first offline homework of the online education training camp submitted by the target student; The text similarity calculation unit, communicatively connected to the answer data extraction unit, is used to calculate the first text similarity between the corresponding standard answer content and the corresponding first answer content for each of the test questions according to the corresponding standard answer content; The listening state classification unit, communicatively connected to the answer data extraction unit, for each of the test questions, intercepts a segment of the in-class listening monitoring video whose acquisition period is consistent with the display period of the lecture page of the knowledge point from the in-class listening monitoring video of the target student according to the corresponding knowledge point, and imports the segment of the in-class listening monitoring video into a pre-trained listening state classification model based on the first machine learning algorithm, and outputs the first confidence level that the target student's listening state for the corresponding knowledge point is a serious listening state; The homework plagiarism classification unit, communicatively connected to the answer data extraction unit, the text similarity calculation unit and the listening state classification unit respectively, is used to import the answer duration, the first text similarity and the first confidence level corresponding to each of the test questions into a pre-trained homework plagiarism classification model based on the second machine learning algorithm, and outputs the second confidence level that there is a plagiarism behavior in the first offline homework of the online education training camp; The judgment unit, communicatively connected to the homework plagiarism classification unit, is configured to judge whether the second confidence level exceeds a preset first confidence level threshold; The determination response unit, communicatively connected to the judgment unit, is configured to, when it is determined that the second confidence level exceeds the first confidence level threshold, re-extract corresponding new test questions for the knowledge points to which the respective test questions belong, and push all the new test questions to the target student as the new offline homework of the target student, and when it is determined that the second confidence level does not exceed the first confidence level threshold, use the submission timestamp of the first online education training camp offline homework as the offline homework check-in timestamp of the target student.

8. A computer device, characterized in that, It includes a memory, a processor, and a transceiver that are communicatively connected in sequence. Among them, the memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer programs and execute the online education training camp offline homework check-in management method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that , and instructions are stored on the computer-readable storage medium. When the instructions are run on a computer, the online education training camp offline homework check-in management method according to any one of claims 1 to 6 is executed.

10. A computer program product, comprising a computer program or instructions, characterized in that, The computer program or the instructions, when executed by a computer, implement the online education training camp offline homework check-in management method according to any one of claims 1 to 6.

Images