A method for automatically generating teaching evaluation text

Abstract

The present invention relates to a method for automatically generating a teaching evaluation text. When in use, a teacher can summarize the required file materials into basic information in the form of folders or documents, and then execute corresponding logical steps to obtain the teaching evaluation text after selectively screening and typeset the contents in the basic information. In this process, a detailed semantic analysis and evaluation are performed on the teaching content in the basic information to obtain a second basic text, which is then filled into a second table document corresponding to a second table label. This method can greatly reduce the workload of teachers for self-editing documents during the process of writing teaching evaluation documents, save editing time, and significantly improve the writing efficiency and writing quality of teaching evaluation texts.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and in particular to a method for automatically generating teaching evaluation texts. Background Art

[0002] Teaching evaluation is based on teaching objectives, uses operational scientific means, systematically collects information about teaching, and makes value judgments on the teaching process and results through quantification to promote the improvement of teaching. At the same time, teaching evaluation texts are also an important part of teaching evaluation. Teaching evaluation texts involve periodic summaries of teachers' current teaching stages. By periodically organizing teaching evaluation texts, it is convenient for teachers to adjust subsequent teaching progress and summarize teachers' personal teaching experience and achievements.

[0003] However, existing education system data reports are often compiled and collected manually, and ultimately produced into various data reports. The operation is very cumbersome, time-consuming and labor-intensive, greatly wasting teachers' precious time, and requiring extraction from various historical data, resulting in low writing efficiency. Summary of the invention

[0004] In view of the above problems, the present invention provides a method for automatically generating a teaching evaluation text, which solves the problem that the existing teaching evaluation text relies on teachers to manually organize and edit data, resulting in low writing efficiency.

[0005] To achieve the above object, the present invention provides a method for automatically generating a teaching evaluation text, comprising:

[0006] Obtain basic information, including teacher information, course information, project information and application information. Teacher information includes teacher name, teacher number, education information, teaching experience information, teacher rating information, teaching achievement information and teaching honor information. Course information includes subject information, planned teaching information of the current course and actual teaching information of the current course. Project information includes the teaching plan information of the current course. Application information includes the application file name and application file number.

[0007] Selecting a preset template in a database according to the declaration information, the preset template includes at least one of declaration form information, typesetting specification information, and data requirement information, the declaration form information includes a form tag and a form column, each form tag corresponds to a form column, the form tag includes a first form tag group and a second form tag group, the first form tag group includes a plurality of first form tags, the second form tag group includes a plurality of second form tags, and the typesetting specification information includes first typesetting specification information and second typesetting specification information;

[0008] Generate a first screening condition according to the first table tag group, perform data screening on the basic information to obtain a first basic text, and fill the first basic text into the table column corresponding to the first table tag according to the first typesetting specification information to obtain a first table document;

[0009] Generate a second screening condition according to the second table tag group, and perform data screening on the basic information to obtain a second basic text, the second basic text including planned teaching content, actual teaching content and teaching objectives;

[0010] Compare the planned teaching content with the actual teaching content to find out the actual teaching problems of the current course;

[0011] Inputting the actual teaching problem into the first neural network model to obtain a solution plan corresponding to the current actual teaching problem, wherein the solution plan includes at least one of a solution suggestion and a solution method for improving or solving the current actual teaching problem;

[0012] Fill the planned teaching content, actual teaching content, teaching objectives, actual teaching problems and solutions into the table column corresponding to the second table label according to the second typesetting specification information to obtain a second table document;

[0013] Generate a final declaration form based on the first form document and the second form document;

[0014] Export the basic information to a file based on the data requirement information and generate attachment information, which includes an attachment list and attached files;

[0015] The final application form and the attachment information are mapped and stored to obtain the teaching evaluation text.

[0016] In some embodiments, generating the first filtering condition according to the first table tag group includes:

[0017] Obtaining the meaning of the first table label, and obtaining the data type corresponding to the first table label according to the meaning of the first table label, where the data type includes at least one of numbers, words, letters, and punctuation marks;

[0018] The meaning of the first table label and the data type are mapped and stored to obtain the current screening index of the first table label;

[0019] Arrange the screening indicators of the plurality of first table tags according to the arrangement order in the first table tag group to obtain a first screening condition;

[0020] The basic information is screened to obtain the first basic text including:

[0021] Convert the first screening condition into an SQL screening sentence segment, and run an application program containing the SQL screening sentence segment on the basic information to obtain a text information value that meets the first screening condition, the text information value having a first initial typesetting format, and sort and merge the text information value to obtain a first basic text;

[0022] Filling the first basic text into the table column corresponding to the first table tag according to the first typesetting specification information, obtaining a first table document including:

[0023] Obtaining a first typesetting specification for each first table label in the first typesetting specification information, wherein the first typesetting specification information includes at least one of font size, word limit, data type, line spacing, alignment, and font type;

[0024] Matching the text information values ​​in the first basic text with the first table tags one by one, and determining whether the current first initial typesetting format complies with the first typesetting specification;

[0025] If yes, fill the text information value into the table column corresponding to the first table tag in the first initial typesetting format;

[0026] If not, a first typesetting processing function model is generated according to the first typesetting specification, and the text information value is input into the first typesetting processing function model for format adjustment, and the format-adjusted text information value is filled into the table column corresponding to the first table label.

[0027] In some embodiments, the second filtering condition is generated according to the second table tag group, and the basic information is filtered to obtain the second basic text including:

[0028] Obtaining the meaning of the second table label and recording it as a screening text topic, and arranging the multiple screening text topics according to the arrangement order in the second table label group to obtain a second screening condition;

[0029] Preprocessing the basic information to obtain basic information containing only paragraphs is recorded as the first intermediate text, and the paragraph in the first intermediate text is recorded as the first paragraph;

[0030] Input the first paragraphs into the semantic analysis model one by one, obtain the paragraph text topics of the first paragraphs, and map and store the paragraph text topics with the first paragraphs;

[0031] Cluster multiple paragraph text topics to obtain several paragraph topic sets, each of which has a central topic;

[0032] The central theme is matched with the screening text theme in the second screening condition in terms of similarity, and the first paragraphs associated with the central theme having the same similarity with the screening text theme are sorted and merged to obtain the second basic text.

[0033] In some embodiments, multiple paragraph text topics are clustered to obtain several paragraph topic sets, each of which has a central topic including:

[0034] Set K first cluster center points, and calculate the first distance value between the paragraph text topic and each first cluster center point one by one;

[0035] According to the first distance value, the paragraph text topic is assigned to the cluster where the first cluster center point closest to it is located, which is recorded as the first cluster cluster;

[0036] Calculate the average vector of all paragraph text topics in each first cluster as the second cluster center of the current first cluster, calculate the second distance value between the paragraph text topic and each second cluster center one by one, and assign the paragraph text topic to the cluster where the second cluster center closest to it is located according to the second distance value, which is recorded as the second cluster;

[0037] Repeat the above steps until the preset number of iterations is reached to obtain multiple final clusters, and the paragraph text topics in each final cluster constitute a paragraph topic set;

[0038] Input the paragraph text topics in the paragraph topic set into the semantic analysis model to obtain the central topic of the current paragraph topic set;

[0039] The central theme is matched with the screening text theme in the second screening condition by similarity, and the first paragraphs associated with the central theme having the same similarity with the screening text theme are sorted and merged, so that the second basic text also includes:

[0040] Converting central topics into central topic vectors, and converting filtered text topics into filtered text topic vectors;

[0041] The cosine similarity algorithm is used to calculate the topic vector of the filtered text and the central topic vector to obtain the cosine similarity between each central topic vector and the topic vector of the filtered text;

[0042] Obtain the similarity threshold corresponding to the current filtered text topic, determine whether the cosine similarity between each central topic and the current filtered text topic is within the range of the similarity threshold, and if so, sort and merge the first paragraph associated with the central topic to obtain the text information of the current filtered text topic, which is recorded as the filtered text information;

[0043] Repeat the above steps until all the filter text topics are traversed to obtain multiple filter text information;

[0044] The multiple filtered text information is sorted and merged to obtain a second basic text.

[0045] In some embodiments, the planned teaching content and the actual teaching content are compared to obtain the actual teaching problems of the current course, including:

[0046] The planned teaching content is divided into the smallest teaching units and arranged in the planned teaching order to obtain a first teaching timeline, and the first teaching timeline is provided with a plurality of first teaching units arranged in sequence; the actual teaching content is divided into the smallest teaching units and arranged in the actual teaching order to obtain a second teaching timeline, and the second teaching timeline is provided with a plurality of second teaching units arranged in sequence;

[0047] Matching the first teaching timeline with the second teaching timeline one by one in chronological order to obtain a plurality of first abnormal events, wherein the first abnormal event includes an abnormal time node and an abnormal event, and the abnormal event includes at least one of a first teaching unit that has not been taught and a second teaching unit that has been newly taught;

[0048] Predicting the causes of the first abnormal events one by one to obtain a plurality of first abnormal causes, wherein the first abnormal cause includes at least one of a student factor, a teacher factor, a classroom factor, and a time factor;

[0049] The first abnormal cause is sorted out to obtain the actual teaching problem.

[0050] In some embodiments, the method further comprises:

[0051] Obtain the planned teaching information in the course information of the next course in the current teaching plan information, and record it as pending teaching information;

[0052] In the solution plan, the teaching units that need to be taught again in the course information of the next course are selected and recorded as abnormal teaching units;

[0053] Determine whether the number of abnormal teaching units exceeds the upper limit of the number of newly added teaching units;

[0054] If yes, then add the course information to the teaching plan information, and generate the newly added course information according to the abnormal teaching unit;

[0055] If not, the abnormal teaching unit is inserted into the pending teaching information and updated to become the scheduled teaching information in the next course information.

[0056] In some embodiments, the method further comprises:

[0057] Perform text recognition on the current actual teaching content to determine whether there is keyword information, which includes at least one of social responsibility, moral concept, civic awareness, ability assessment, social response, and practical results;

[0058] If there is keyword information, the actual teaching content is input into the second neural network model to obtain a follow-up investigation plan corresponding to the actual teaching content, the follow-up investigation plan including at least one of the investigation method, investigation content, and tracking duration and tracking frequency;

[0059] Fill the follow-up survey plan into the table column of the second table tag associated with the follow-up survey plan, or add the follow-up survey plan as an attachment to the attachment information.

[0060] In some embodiments, the planned teaching content, the actual teaching content, the teaching objectives, the actual teaching problems and the solutions are recorded as the text to be processed, and the planned teaching content, the actual teaching content, the teaching objectives, the actual teaching problems and the solutions are filled into the table column corresponding to the second table label according to the second typesetting specification information, and further includes:

[0061] Obtaining the second typesetting specification for each second table label in the second typesetting specification information, wherein the second typesetting specification information includes at least one of font size, word limit, line spacing, alignment, and font type;

[0062] Matching the text to be processed with the second table tags one by one, wherein the text to be processed has a second initial typesetting format;

[0063] Determine one by one whether the current second initial typesetting format complies with the second typesetting specification;

[0064] If yes, fill the text to be processed into the table column corresponding to the second table label in the second initial typesetting format;

[0065] If not, a second typesetting processing function model is generated according to the second typesetting specification, and the text to be processed is input into the second typesetting processing function model for format adjustment to obtain a first processed text;

[0066] Determine whether the word count of the first processed text meets the second typesetting specification, and if not, input the content of the first processed text into the third neural network model to obtain a second processed text with a limited word count that meets the second typesetting specification;

[0067] The first processed text and the second processed text are mapped and stored, and the second processed text is filled into the table column corresponding to the second table label, and the first processed text is displayed in the sidebar of the table column of the current second table label in the form of annotation.

[0068] In some embodiments, the method further comprises:

[0069] Determine whether the number of annotations in the current second table document exceeds a preset annotation number threshold;

[0070] If yes, the second table labels containing the annotations are sorted and an annotation index is generated;

[0071] Displaying the annotation index in a sidebar of the second table document;

[0072] Obtain the input command in the sidebar and identify the input coordinates of the input command;

[0073] Match the second table tag within the current annotation index according to the typed coordinates;

[0074] The first processing text associated with the second table tag is displayed in a pop-up window.

[0075] In some embodiments, the first neural network model is trained in the following manner:

[0076] Constructing a first convolutional neural network model, and extracting historical teaching problems and historical solutions from a database to generate a first sample database, wherein first sample data in the first sample database includes sample teaching problems and sample solutions;

[0077] Selecting first sample data to train a first convolutional neural network model until the accuracy of the test solution output by the first convolutional neural network model and the sample solution reaches a first preset threshold, thereby obtaining a first neural network model;

[0078] The second neural network model is trained in the following way:

[0079] Constructing a second convolutional neural network model, and extracting historical actual teaching content and historical tracking survey plans from the database to generate a second sample database, wherein the second sample data in the second sample database includes sample actual teaching content and sample tracking survey plans;

[0080] Selecting second sample data to train the second convolutional neural network model until the accuracy of the test tracking survey plan and the sample tracking survey plan output by the second convolutional neural network model reaches a second preset threshold, thereby obtaining a second neural network model;

[0081] The third neural network model is trained in the following way:

[0082] Constructing a third convolutional neural network model, and extracting historical processing text and historical summary text from the database to generate a third sample database, wherein the third sample data in the third sample database includes the sample processing text and the sample summary text;

[0083] A third sample data is selected to train the third convolutional neural network model until the accuracy of the test summary text and the sample summary text output by the third convolutional neural network model reaches a third preset threshold, thereby obtaining a third neural network model.

[0084] Different from the prior art, the above technical solution obtains basic information, which includes teacher information, course information, project information and application information, and selects a preset template in the database according to the application information, generates a first screening condition according to the first table tag group, and performs data screening on the basic information to obtain a first basic text, and fills the first basic text into the table column corresponding to the first table tag according to the first typesetting specification information to obtain a first table document; generates a second screening condition according to the second table tag group, and performs data screening on the basic information to obtain a second basic text, the second basic text includes planned teaching content, actual teaching content and teaching objectives; compares the differences between the planned teaching content and the actual teaching content to obtain Actual teaching problems of the current course; inputting the actual teaching problems into the first neural network model to obtain solutions corresponding to the current actual teaching problems, the solutions including at least one of the suggestions for improving or solving the current actual teaching problems and solutions; filling the planned teaching content, actual teaching content, teaching objectives, actual teaching problems and solutions into the table column corresponding to the second table label according to the second typesetting specification information to obtain a second table document; generating a final application form based on the first table document and the second table document; exporting the basic information to a file according to the data requirement information and generating attachment information, the attachment information including an attachment list and attached files; mapping and storing the final application form and the attachment information to obtain a teaching evaluation text. Teachers can summarize the required file materials into basic information in the form of folders or documents, and then execute the corresponding logical steps to obtain the teaching evaluation text after selectively screening and typeset the content in the basic information. In this process, the teaching content in the basic information is carefully semantically analyzed and evaluated to obtain the second basic text, which is then filled in the second table document corresponding to the second table label. This method can greatly reduce the workload of teachers' independent editing of documents during the writing process of teaching evaluation documents, save editing time, and significantly improve the writing efficiency and quality of teaching evaluation texts.

[0085] The above-mentioned records related to the content of the invention are only an overview of the technical solution of the present invention. In order to enable ordinary technicians in the field to more clearly understand the technical solution of the present invention, and then implement it according to the text of the specification and the contents recorded in the drawings, and in order to make the above-mentioned purpose and other purposes, features and advantages of the present invention easier to understand, the following is a description in combination with the specific implementation mode of the present invention and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] The drawings are only used to illustrate the principles, implementations, applications, features and effects of the specific embodiments of the present invention and other related contents, and shall not be considered as limiting the present invention.

[0087] In the drawings of the specification:

[0088] Figure 1 It is a schematic diagram of step S101 to step S110 of the generation method described in the specific embodiment of the present invention;

[0089] Figure 2 It is a schematic diagram of steps S201 to S208 of the generation method described in the specific implementation manner of the present invention;

[0090] Figure 3 It is a schematic diagram of steps S301 to S305 of the generation method described in the specific implementation manner of the present invention;

[0091] Figure 4 It is a schematic diagram of step S401 to step S404 of the generation method described in the specific implementation manner of the present invention;

[0092] Figure 5 It is a schematic diagram of steps S501 to S509 of the generation method described in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0093] In order to explain in detail the possible application scenarios, technical principles, specific schemes that can be implemented, and the purposes and effects of the present invention, the following is a detailed description in combination with the specific embodiments listed and the accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of the present invention, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present invention.

[0094] Reference to "embodiment" herein means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The word "embodiment" appearing in various places in the specification does not necessarily refer to the same embodiment, nor does it particularly limit its independence or association with other embodiments. In principle, in the present invention, as long as there is no technical contradiction or conflict, the various technical features mentioned in each embodiment can be combined in any way to form a corresponding implementable technical solution.

[0095] Unless otherwise defined, the technical terms used herein have the same meanings as those generally understood by those skilled in the art to which the present invention belongs; the use of relevant terms herein is only for describing specific embodiments and is not intended to limit the present invention.

[0096] In the description of the present invention, the term "and / or" is an expression used to describe the logical relationship between objects, indicating that three relationships may exist, for example, A and / or B, which means: A exists, B exists, and A and B exist at the same time. In addition, the character " / " in this article generally indicates that the objects before and after are in a logical relationship of "or".

[0097] In the present invention, terms such as "first" and "second" are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship of quantity, priority or sequence between these entities or operations.

[0098] Without further restrictions, in the present invention, the words "include", "comprises", "has" or other similar open-ended expressions used in sentences are intended to cover non-exclusive inclusion. These expressions do not exclude the presence of additional elements in the process, method or product including the elements, so that the process, method or product including a series of elements may include not only those limited elements, but also other elements not explicitly listed, or also include elements inherent to such process, method or product.

[0099] Similar to the understanding in the Examination Guidelines, in the present invention, expressions such as "greater than", "less than", "exceed" and the like are understood to exclude the number itself; expressions such as "above", "below", "within" and the like are understood to include the number itself. In addition, in the description of the embodiments of the present invention, "multiple" means more than two (including two), and similar expressions related to "multiple" are also understood in this way, such as "multiple groups", "multiple times", etc., unless otherwise clearly and specifically limited.

[0100] See also Figure 1 This embodiment provides a method for automatically generating a teaching evaluation text, including:

[0101] S101, obtaining basic information, which includes teacher information, course information, project information and application information. Teacher information includes teacher name, teacher number, education information, teaching experience information, teacher rating information, teaching achievement information and teaching honor information. Course information includes subject information, planned teaching information of the current course and actual teaching information of the current course. Project information includes teaching plan information of the current course. Application information includes application file name and application file number.

[0102] S102, selecting a preset template in a database according to the declaration information, the preset template including at least one of declaration form information, typesetting specification information, and data requirement information, the declaration form information including a form tag and a form column, each form tag corresponding to a form column, the form tag including a first form tag group and a second form tag group, the first form tag group including a plurality of first form tags, the second form tag group including a plurality of second form tags, the typesetting specification information including first typesetting specification information and second typesetting specification information;

[0103] S103, generating a first screening condition according to the first table tag group, and performing data screening on the basic information to obtain a first basic text, and filling the first basic text into a table column corresponding to the first table tag according to the first typesetting specification information to obtain a first table document;

[0104] S104, generating a second screening condition according to the second table tag group, and performing data screening on the basic information to obtain a second basic text, the second basic text including planned teaching content, actual teaching content and teaching objectives;

[0105] S105, comparing the planned teaching content with the actual teaching content to obtain actual teaching problems of the current course;

[0106] S106, inputting the actual teaching problem into the first neural network model to obtain a solution plan corresponding to the current actual teaching problem, wherein the solution plan includes at least one of a solution suggestion and a solution method for improving or solving the current actual teaching problem;

[0107] S107, filling the planned teaching content, actual teaching content, teaching objectives, actual teaching problems and solutions into the table column corresponding to the second table label according to the second typesetting specification information to obtain a second table document;

[0108] S108, generating a final declaration form according to the first form document and the second form document;

[0109] S109, export the basic information to a file according to the data requirement information and generate attachment information, the attachment information including an attachment list and attached files;

[0110] S110, mapping and storing the final application form and the attachment information to obtain a teaching evaluation text.

[0111] In step S101, the basic information can be the basic materials that the user has simply sorted out. The embodiment of the basic information can be a file in a variety of formats or a separate document information, which is not limited in this embodiment. For the convenience of expression, the basic information is divided into teacher information, course information, project information and declaration information according to the materials and content required to fill in the teaching evaluation text, wherein the teacher information includes the teacher's name, teacher number, academic information, teaching experience information, teacher rating information, teaching achievement information and teaching honor information, wherein the teacher number can be the current user's electronic ID number or the number with unique identification information used in the current teaching system. Academic information is the teacher's personal academic qualifications, such as bachelor, master, doctor, etc., teaching experience information includes the experience of the current user since he / she has been engaged in teaching work, and teacher rating information is the official certified grade information obtained by the current user in the current teaching system or education assessment indicators, such as senior teacher title, special teacher title, etc. Teaching achievement information can be understood as the interim results obtained by the teaching project that the current user is responsible for since he / she has been engaged in teaching work, and teaching honor information is the honors obtained by the current user since he / she has been engaged in teaching work. It can be understood that if the current user is a teaching team, the current teacher information includes all or part of the teacher information in the current teaching team.

[0112] In this embodiment, the course information is the course information of the current teaching, and can also be understood as the course body of the teaching evaluation required by the current user. It should be noted that the teaching evaluation text is the material after the teaching is completed, which belongs to the reprocessing of the course materials and the data generated by the teaching process after class. On this basis, the course information includes actual teaching information and planned teaching information. The planned teaching information can be subdivided into the course materials, course lesson plans and course plans prepared by the current user, the knowledge points mastered by the course plan, etc. The actual teaching information is the data collected during the actual teaching process. The actual teaching information may include the teacher's teaching video, teaching voice, the teaching interaction data between students and teachers, the knowledge points of the actual teaching, etc.

[0113] In this embodiment, the project information is the information of the teaching project to which the course information that the current user is responsible belongs. The project information includes the teaching plan information described in the current course. For example, if a course is taught for College English Level 4, the project information includes the teaching plan information for College English Level 4 for the current four years of the university. For another example, if the teaching project that the current user is responsible for is advanced physics teaching in colleges and universities, the project information is the physics teaching technology plan formulated for colleges and universities, and the course information is the course information of a certain section in physics teaching, or the physics teaching course information of a certain semester.

[0114] In this embodiment, the declaration information includes the declaration file name and the declaration file number. The declaration file name is the file name corresponding to the current teaching evaluation text, such as the interim teaching evaluation form for the xth semester of the xth grade of xx University; the declaration file number is the number of the declaration file in the teaching system.

[0115] In step S102, a preset template can be selected from the database based on the declaration information. It should be noted that the database can be understood as a teaching database of a college or university, or it can be a separately constructed database. In this database, there are preset templates with multiple declaration file numbers, as well as teaching evaluation text information written by each user. The preset template corresponding to this declaration file is screened out in the database according to the declaration file number. Specifically, the preset template includes declaration form information, typesetting specification information, and data requirement information. Among them, the declaration form information is a form filled in the teaching evaluation text. For ease of expression, the declaration form is split into a table label and a table column. The table label is the title of each table column. For example, if the first line of the declaration form is the name of the form, then the table label is "Name" and the table column is the text box in the first line.

[0116] Furthermore, the table tags are divided into a first table tag group and a second table tag group according to the filling requirements. The first table tags are table tags that do not require text semantic processing but only require typeset information organization, such as "name, experience, ID number, contact number, education, gender", etc. The first table tags can also be understood as table tags that are not strongly related to the current course subject, such as "past teaching achievements, past teaching honors, teacher's personal achievements", etc.; the second table tags are table tags that require text semantic processing and typeset information organization after processing, such as "course objectives, course plans, actual teaching content, course problems, course improvement plans, course prospects, course evaluations", etc. The second table tags can also be understood as table tags that are more closely related to the course subject of the current teaching evaluation text.

[0117] In this embodiment, the typesetting specification information is divided into first typesetting specification information and second typesetting specification information, the first typesetting specification information is the typesetting format requirement corresponding to the first table tag, and the second typesetting specification information is the typesetting format requirement corresponding to the second table tag. In this embodiment, the material requirement information can be understood as the requirement information of the materials that need to exist separately in the form of files during the process of filling in the teaching evaluation text, including the name of the materials.

[0118] In step S103, the first filtering condition is generated according to the first table tag group, and the basic information is screened according to the first filtering condition to obtain the first basic text. For example, if the first filtering condition contains "name", the teacher's name is filtered in the basic information and organized into the first basic text. It should be noted that the first filtering condition includes the filtering conditions of multiple table tags in the first table tag group, as described below. After obtaining the first basic text, the first basic text is written into the table column corresponding to each first table tag according to the first typesetting specification information to obtain the first table document.

[0119] In step S104, a second filtering condition is generated according to the second table tag group, and the basic information is filtered under the second filtering condition to obtain a second basic text. It can be understood that the second basic text is the course materials associated with the current course body, specifically including planned teaching content, actual teaching content and teaching objectives.

[0120] In step S105, the planned teaching content is compared with the actual teaching content. Specifically, the planned teaching content is the preliminary preparation made by the current user for this course and a reasonable prediction of the course teaching process. In the actual teaching process, some subjective or objective factors may occur, such as the discussion time in class exceeds the expected time; the student speech time in class is too long; the teaching equipment in class has a malfunction and needs temporary repair; the knowledge points taught in class cannot meet expectations, and the teaching time is extended, etc., resulting in differences between the actual teaching content and the planned teaching content. This embodiment compares the planned teaching content with the actual teaching content to obtain the actual teaching problem of the current course, which can also be understood as the cause and result of the difference between the actual teaching content and the planned teaching content.

[0121] In step S106, the actual teaching problem is input into the first neural network model, and the first neural network model gives a corresponding solution plan for the actual teaching problem. Specifically, the solution plan may be a solution suggestion or a solution. For example, if the discussion time in class exceeds the expected time, resulting in the omission of teaching knowledge points, the solution plan may be the way to add the currently omitted teaching knowledge points in the next course and the limit on the teaching time, so that the current user can make timely adjustments in the teaching plan of the next course.

[0122] In step S107, the planned teaching content, actual teaching content, teaching objectives, actual teaching problems and solutions are filled into the table column corresponding to the second table label according to the second typesetting specification information to obtain the second table document, thereby realizing the automatic generation function of the second table document.

[0123] In step S108, the first table document and the second table document are merged to obtain the declaration form. The specific merging process may be to match the table tags of the declaration form in the preset template with the table tags in the current first table document and the second table document one by one, and adjust the arrangement order of the table columns in the first table document and the second table document according to the table tag sequence of the declaration form in the preset template to obtain the final declaration form.

[0124] In step S109, the basic information is exported according to the data requirement information, which may be specifically: obtaining the data name in the data requirement information, and constructing a folder corresponding to each data name; identifying the file name in the basic information, matching the file name with the data name, and merging the file documents with the successfully matched file name into the folder corresponding to the data name, and presenting the data name in a list to form an attachment list.

[0125] In step S110, the final declaration form and the attachment information are mapped and stored to obtain a teaching evaluation text.

[0126] In this embodiment, the teacher can preliminarily organize or simply summarize the required file materials in the form of folders or documents to form basic information. After executing the logical steps shown in this embodiment, the content in the basic information can be selectively screened and typeset to obtain the teaching evaluation text. In this process, the teaching content in the basic information is carefully semantically analyzed and evaluated to obtain the second basic text. At the same time, the actual teaching content and the planned teaching content are compared to obtain the actual teaching problems. A solution is generated for the actual teaching problem using a deep learning algorithm, and the current user is given teaching suggestions or teaching plan improvements, which are then filled in the second table document corresponding to the second table label. This method can greatly reduce the workload of teachers for self-editing documents during the writing process of teaching evaluation documents, save editing time, and significantly improve the writing efficiency and quality of teaching evaluation texts, so that teachers can focus on the text content in the second table document in this process, and manually adjust and modify the text content in the second table document, so as to break away from the blind and cumbersome steps of arranging materials and preliminary writing, and focus on the editing and rewriting of the central document of the current teaching evaluation text, improve the writing quality of the teaching evaluation text, and reduce the writing burden of teachers.

[0127] See also Figure 2 In some embodiments, generating the first screening condition according to the first table tag group includes:

[0128] S201, obtaining the meaning of a first table label, and obtaining a data type corresponding to the first table label according to the meaning of the first table label, where the data type includes at least one of numbers, words, letters, and punctuation marks;

[0129] S202, mapping and storing the meaning of the first table label and the data type, and obtaining the current screening index of the first table label;

[0130] S203, sorting the screening indicators of the plurality of first table tags according to the arrangement order in the first table tag group to obtain a first screening condition;

[0131] The basic information is screened to obtain the first basic text including:

[0132] S204, converting the first screening condition into an SQL screening sentence segment, and running an application program containing the SQL screening sentence segment on the basic information to obtain a text information value that meets the first screening condition, the text information value having a first initial typesetting format, and arranging and merging the text information value to obtain a first basic text;

[0133] Filling the first basic text into the table column corresponding to the first table tag according to the first typesetting specification information, obtaining a first table document including:

[0134] S205, obtaining a first typesetting specification for each first table label in the first typesetting specification information, wherein the first typesetting specification information includes at least one of font size, word limit, data type, line spacing, alignment, and font type;

[0135] S206, matching the text information values ​​in the first basic text with the first table tags one by one, and determining whether the current first initial typesetting format complies with the first typesetting specification;

[0136] S207: If yes, fill the text information value into the table column corresponding to the first table tag in the first initial typesetting format;

[0137] S208. If not, generate a first typesetting processing function model according to the first typesetting specification, input the text information value into the first typesetting processing function model for format adjustment, and fill the format-adjusted text information value into the table column corresponding to the first table tag.

[0138] In step S201, the meaning of the first table label is obtained. The meaning of the first table label can be obtained through text recognition or semantic recognition, or it can be manually written into the preset template during the generation stage of the preset template. The different meanings of the first table label correspond to different data types. The data type can be understood as the conventional embodiment of the first table label. For example, when the first table label is "contact number", the meaning of the first table label is a mobile phone number or a landline number, and the data type of the first table label corresponds to Arabic numerals; for another example, when the first table label is "email", the meaning of the first table label is an email address, and the data type of the first table label corresponds to a combination of numbers, letters and symbols.

[0139] In step S202, the meaning of the first table tag is mapped with the data type and stored to obtain the screening index of the first table tag. It is understandable that the screening index can be embodied in the form of a character string. This method is convenient for narrowing the screening range of each first table tag in the subsequent screening process. Taking "contact number" as an example, the screening index of the first table tag is an 11-digit Arabic numeral combination, which can exclude characters that are not 11-digit Arabic numeral combinations in the basic information, thereby achieving fast matching and screening of the first table tag. Steps S201 and S202 can be repeated according to the number of first table tags in the first table tag group until all first table tags are identified, forming screening indicators for multiple first table tags.

[0140] In step S203, the screening indicators of the plurality of first table tags are sorted according to the arrangement order in the first table tag group to obtain a first screening condition.

[0141] In step S204, the SQL language can be selected to compile the first filtering condition, and the SQL application program can be run in the text composed of the basic information, so as to implement the filtering in the basic information and obtain the text information value that meets the first filtering condition. The text information value can be text, numbers, English, characters, etc. In this embodiment, the text information value has a first initial typesetting format, such as the default text format is Songti 5, the English format is Roman font 5, etc., which can be set according to actual needs, and the filtered text information values ​​are sorted to form the first basic text. This sorting process can be directly integrated in the SQL application program. It should be noted that each text information value in the first basic text corresponds to a filtering indicator. In step S206, the filtering indicator corresponding to each text information value is matched with the first table label.

[0142] In step S205, the first typesetting specification in the first typesetting specification information is obtained, and in step S206, each text information value is matched, and after the match is successful, it is determined whether the current first initial typesetting format conforms to the first typesetting specification.

[0143] In step S207, if the current first initial typesetting format complies with the first typesetting specification, the text information value may be directly written into the table column corresponding to the first table tag to which it is associated.

[0144] In step S208, if the current first initial typesetting format does not conform to the first typesetting specification, it is necessary to generate a first typesetting processing function model according to the first typesetting specification. Specifically, the first typesetting specification can be compiled using a programming language dedicated to text format conversion to form a first typesetting processing function model, which can be Python. Further, the first typesetting processing function model is used to adjust the format of each text information value, and the text information value after the format adjustment is filled into the table column corresponding to the first table label.

[0145] This embodiment shows the generation process of the first screening condition and the specific steps of obtaining the first form document based on the first basic text. The technical solution shown in this embodiment can save users the steps of filling in some routine information, including copying and pasting and formatting of form-filling information such as name, ID card, contact number, email address, education level, etc., saving writing time and improving writing efficiency.

[0146] See also Figure 3 In some embodiments, the second filtering condition is generated according to the second table tag group, and the basic information is filtered to obtain the second basic text including:

[0147] S301, obtaining the meaning of the second table label and recording it as a screening text subject, and arranging multiple screening text subjects according to the arrangement order in the second table label group to obtain a second screening condition;

[0148] S302, preprocessing the basic information to obtain basic information containing only paragraphs, which is recorded as the first intermediate text, and the paragraph in the first intermediate text is recorded as the first paragraph;

[0149] S303, inputting the first paragraphs into the semantic analysis model one by one, obtaining the paragraph text topics of the first paragraphs, and mapping and storing the paragraph text topics and the first paragraphs;

[0150] S304, clustering the multiple paragraph text topics to obtain a number of paragraph topic sets, each paragraph topic set having a central topic;

[0151] S305: perform similarity matching between the central theme and the screening text theme in the second screening condition, sort out and merge the first paragraphs associated with the central theme having the same similarity with the screening text theme, and obtain a second basic text.

[0152] In this embodiment, the content corresponding to the second table tag is a paragraph text or a combination of words with a certain number of words, which is different from the first table tag, so the generation of the second filtering condition is also different from the generation method of the first filtering condition. In step S301, the meaning of the second table tag is used as the filtering text topic, and each second table tag is a filtering text topic. Then, the filtering text topics are sorted according to the arrangement order in the second table tag group to obtain the second filtering condition. Optionally, the filtering text topics and the second table tags can be mapped and stored and then arranged in sequence.

[0153] In step S302, the basic information is preprocessed, and the preprocessing includes text data cleaning, word segmentation processing, and stop word processing, etc., wherein text data cleaning can remove unnecessary special characters, numbers, etc. in the basic information, and word segmentation processing includes dividing the text information based on punctuation marks, specifically including identifying multiple paragraphs in the basic information based on paragraph symbols, and identifying each sentence based on periods, and can also include word segmentation for each sentence, etc. In this embodiment, the basic information containing paragraphs can be obtained through preprocessing, and it is recorded as the first intermediate text, and the paragraph contained in the first intermediate text is recorded as the first paragraph.

[0154] In step S303, the semantic analysis model may be an LDA algorithm model or an LSA algorithm model, and the first paragraphs are input into the semantic analysis model one by one to obtain the paragraph text topic corresponding to the first paragraph text.

[0155] In step S304, multiple paragraph text topics are classified and aggregated according to their similarity or the closeness of semantic association to obtain multiple paragraph topic sets, each of which contains at least one paragraph text topic, and a central topic is generated for each paragraph topic set. In this embodiment, the central topic is obtained by further refining the semantics represented by multiple paragraph text topics. For example, a feasible central topic generation method may be to concatenate paragraph text topics in the same paragraph topic set with punctuation marks to form a paragraph text, recorded as the second paragraph, and input the second paragraph into the semantic analysis model to obtain the text topic of the second paragraph, that is, the central topic.

[0156] In step S305, the central theme is matched with the filtered text theme in the second filtering condition for similarity. If the similarity between the central theme and the filtered text theme is the same, a paragraph theme set corresponding to the central theme is obtained, and the first paragraph associated with the paragraph text theme in the paragraph theme set is merged to obtain the second basic text.

[0157] Furthermore, in some embodiments, multiple paragraph text topics are clustered to obtain several paragraph topic sets, each of which has a central topic including:

[0158] Set K first cluster center points, and calculate the first distance value between the paragraph text topic and each first cluster center point one by one;

[0159] According to the first distance value, the paragraph text topic is assigned to the cluster where the first cluster center point closest to it is located, which is recorded as the first cluster cluster;

[0160] Calculate the average vector of all paragraph text topics in each first cluster as the second cluster center of the current first cluster, calculate the second distance value between the paragraph text topic and each second cluster center one by one, and assign the paragraph text topic to the cluster where the second cluster center closest to it is located according to the second distance value, which is recorded as the second cluster;

[0161] Repeat the above steps until the preset number of iterations is reached to obtain multiple final clusters, and the paragraph text topics in each final cluster constitute a paragraph topic set;

[0162] Input the paragraph text topics in the paragraph topic set into the semantic analysis model to obtain the central topic of the current paragraph topic set;

[0163] The central theme is matched with the screening text theme in the second screening condition by similarity, and the first paragraphs associated with the central theme having the same similarity with the screening text theme are sorted and merged, so that the second basic text also includes:

[0164] Converting central topics into central topic vectors, and converting filtered text topics into filtered text topic vectors;

[0165] The cosine similarity algorithm is used to calculate the topic vector of the filtered text and the central topic vector to obtain the cosine similarity between each central topic vector and the topic vector of the filtered text;

[0166] Obtain the similarity threshold corresponding to the current filtered text topic, determine whether the cosine similarity between each central topic and the current filtered text topic is within the range of the similarity threshold, and if so, sort and merge the first paragraph associated with the central topic to obtain the text information of the current filtered text topic, which is recorded as the filtered text information;

[0167] Repeat the above steps until all the filter text topics are traversed to obtain multiple filter text information;

[0168] The multiple filtered text information is sorted and merged to obtain a second basic text.

[0169] In this embodiment, the topic clustering method can adopt the K-means clustering algorithm, and the K value can be reasonably set, for example, the number of filtered text topics is obtained, and the number of filtered text topics is used as the K value. On this basis, the K-means clustering algorithm is performed, and the number of the final central topics is K.

[0170] It should be noted that the K-means clustering algorithm is an unsupervised classification algorithm. Specifically, before classification, K first cluster centers are set as the initial cluster centers, and the first distance value of each paragraph text theme from each first cluster center is calculated one by one, and the paragraph text theme is classified into the cluster where the first cluster center with the closest distance value is located according to the first distance value, which is recorded as the first cluster cluster. It should be noted that the first distance value can be calculated by Euclidean distance. Taking the calculation of a single paragraph text theme and a single first cluster center as an example, the current paragraph text theme is converted into a vector form, which is recorded as the first vector, and the current first cluster center is converted into a vector form, which is recorded as the second vector. The distance between the first vector and the second vector is calculated by using the Euclidean distance formula, that is, the first distance value between the first vector and the second vector is obtained. The size of the first distance value is used to judge the distance between each paragraph text theme and the first cluster center, and the paragraph text theme is assigned to the cluster where the first cluster center with the closest distance is located, forming a total of K first cluster clusters.

[0171] When the first cluster is formed, it is recorded as a classification. On this basis, the average vector of all paragraph text themes in each first cluster is calculated, and it is used as the second cluster center point in the first cluster, forming a total of K second cluster centers. On this basis, the distance value between each paragraph text theme and each second cluster center point is continued to be calculated, which is recorded as the second distance value. Similarly, the distance between each paragraph text theme and the second cluster center point is judged according to the second distance value, and the paragraph text theme is assigned to the cluster where the second cluster center point closest to it is located, which is recorded as the second cluster cluster, and recorded as a secondary classification. By analogy, the calculation of the third cluster center point is performed on the basis of the second cluster cluster to obtain the third cluster cluster, which is recorded as three classifications... until the preset number of iterations is reached. It can be understood that the preset number of iterations can be set according to actual needs. After reaching the preset number of iterations, stop classifying the paragraph text theme, and record the cluster at this time as the final cluster cluster, and embody the paragraph text theme in the same final cluster cluster in the form of a set, that is, obtain a paragraph theme set.

[0172] Furthermore, the central theme is matched with the screening text theme in the second screening condition for similarity. Specifically, the cosine similarity algorithm can be used to calculate the similarity of the two texts, and the similarity threshold can be set according to actual needs. On the basis of meeting the similarity threshold, the first paragraph associated with the central theme is sorted and merged to obtain the text information corresponding to the current screening text theme, which is recorded as the screening text information. It should be noted that the text information of different screening text themes can have some or all of the same content. For example, the actual teaching content and the planned teaching content can belong to the same screening text theme of course content, and the actual teaching content can also belong to the screening text theme of practical content. This method can achieve complete aggregation and segmentation of the screening text information.

[0173] After the plurality of filtered text information and the second table label are mapped and stored, the second basic text is obtained.

[0174] This embodiment shows the specific process of obtaining the second screening range based on the second table tag and processing the basic information under the second screening range to obtain the second basic text, which is converted from the user's autonomous writing mode to the use of algorithms to perform certain logical analysis and evaluation on the connotation of the basic information to form a text, so that the user can make manual modifications and adjustments on this basis. It realizes regular arrangement and aggregation based on the semantic content in the basic information, saves the user from manual thought extraction and analysis during the writing process, saves writing steps, and improves writing efficiency.

[0175] See also Figure 4 In some embodiments, the planned teaching content and the actual teaching content are compared to obtain the actual teaching problems of the current course, including:

[0176] S401, dividing the planned teaching content according to the minimum teaching unit, and arranging them according to the planned teaching sequence, to obtain a first teaching timeline, wherein the first teaching timeline is provided with a plurality of first teaching units arranged in sequence, dividing the actual teaching content according to the minimum teaching unit, and arranging them according to the actual teaching sequence, to obtain a second teaching timeline, wherein the second teaching timeline is provided with a plurality of second teaching units arranged in sequence;

[0177] S402, matching the first teaching timeline with the second teaching timeline one by one in chronological order to obtain a plurality of first abnormal events, wherein the first abnormal event includes an abnormal time node and an abnormal event, and the abnormal event includes at least one of a first teaching unit that has not been taught and a second teaching unit that has been newly taught;

[0178] S403, predicting the causes of the first abnormal events one by one to obtain multiple first abnormal causes, where the first abnormal causes include at least one of a student factor, a teacher factor, a classroom factor, and a time factor;

[0179] S404. Arrange the first abnormal cause to obtain the actual teaching problem.

[0180] In step S401, the minimum teaching unit can be understood as the teaching content involved in the minimum knowledge point split during the course teaching process. The planned teaching content in a single course information can contain multiple minimum knowledge points, thus forming a complete course. Based on this idea, when comparing the differences between the planned teaching content and the actual teaching content, the planned teaching content is divided according to the minimum teaching unit to obtain a first teaching timeline and multiple first teaching units arranged in the planned teaching order on the first teaching timeline; and the actual teaching content is divided according to the minimum teaching unit to obtain a second teaching timeline and multiple second teaching units arranged in the actual teaching order on the second teaching timeline.

[0181] In step S402, the first teaching timeline and the second teaching timeline are matched one by one in chronological order to determine whether the first teaching unit and the second teaching unit at the same time node are the same. If not, the current first teaching unit is recorded as the second abnormal event, and the second abnormal event is matched with the remaining second teaching units on the second teaching timeline to determine whether the current second abnormal event has completed the teaching in the subsequent teaching process. If the teaching content of the current second abnormal event still cannot be found in the remaining second teaching units on the second teaching timeline, the current second abnormal event is updated to the first abnormal event, and the time node where the first teaching unit and the second teaching unit are different is recorded as the abnormal event node, and the status of the first teaching unit is recorded as not taught, and an abnormal event is obtained.

[0182] Furthermore, after the comparison in chronological order is completed, the remaining second teaching units that do not match the first teaching unit are the teaching contents newly added during the actual teaching process, and also belong to the first abnormal event. Based on this, the newly added second teaching unit is recorded as a first abnormal event, and the time node of this second teaching unit is recorded as an abnormal time node.

[0183] In step S403, the causes of the first abnormal event are predicted one by one. The specific prediction algorithm can be obtained with the help of a deep learning algorithm model. For example, a basic deep learning algorithm model can be constructed with the help of an MLP algorithm, and the basic deep learning algorithm model can be trained to have the function of predicting the cause of the teaching based on the teaching content. The first abnormal cause can specifically include at least one of student factors, teacher factors, classroom factors, and time factors. Furthermore, student factors can be further subdivided, including student understanding, student application, etc. Teacher factors include the degree of control of the teacher's classroom teaching rhythm, the teacher's teaching habits, the teacher's teaching speed, etc. Classroom factors can include classroom activity, classroom attendance, classroom participation, etc. Time factors can include student class time, teacher class time, student attendance, etc.

[0184] In step S404, the first abnormal causes are sorted to obtain actual teaching problems. Specifically, if there are multiple first abnormal causes, the multiple first abnormal causes can be textually concatenated to form actual teaching problems.

[0185] This embodiment shows the specific process of comparing the actual teaching content with the planned teaching content. This method can facilitate the current user to check for omissions and make up for the gaps in the completed courses, and adjust and add the first teaching unit that was not taught in the first abnormal event to the next course or subsequent teaching plan information, so as to realize comprehensive monitoring and management of knowledge points of teaching courses, realize dynamic adjustment of teaching plan information, improve subsequent teaching quality, and give teaching evaluation texts a deeper meaning.

[0186] In some embodiments, the method further comprises:

[0187] Obtain the planned teaching information in the course information of the next course in the current teaching plan information, and record it as pending teaching information;

[0188] In the solution plan, the teaching units that need to be taught again in the course information of the next course are selected and recorded as abnormal teaching units;

[0189] Determine whether the number of abnormal teaching units exceeds the upper limit of the number of newly added teaching units;

[0190] If yes, then add the course information to the teaching plan information, and generate the newly added course information according to the abnormal teaching unit;

[0191] If not, the abnormal teaching unit is inserted into the pending teaching information and updated to become the scheduled teaching information in the next course information.

[0192] In this embodiment, the current teaching plan information records the course information of the next course, and the course information of the next course contains planned teaching information. For the convenience of expression, it is recorded as pending teaching information. The abnormal teaching unit can be understood as the first teaching unit that is not taught in the actual teaching content of the current course. Further, the number of abnormal teaching units is determined, that is, the number of first teaching units that are not taught in the actual teaching content of the current course is determined. This judgment step facilitates the control of the teaching rhythm in the next course.

[0193] Specifically, if the number of abnormal teaching units is too large, the teaching time of the next course will be too saturated, which will reduce the teaching efficiency and make the planned teaching information of the next course unable to be completed. On this basis, this embodiment controls the number of abnormal teaching units, sets the upper limit of the number of newly added teaching units, and each planned teaching information has an upper limit of the number of newly added teaching units. The specific value of the upper limit of the number of newly added teaching units can be reasonably adjusted according to the planned teaching information and the teaching time of a single course. If the number of abnormal teaching units does not exceed the upper limit of the number of newly added teaching units, the abnormal teaching units can be inserted into the current pending teaching information and updated to the established planned teaching information, that is, as the final planned teaching information of the next course. If the number of abnormal teaching units exceeds the upper limit of the number of newly added teaching units, it means that there are many abnormal teaching units in the current course, and an additional course needs to be added to conduct teaching of abnormal teaching units, that is, add new course information in the teaching plan, and generate new planned teaching information of the newly added course information according to the abnormal teaching units.

[0194] This embodiment shows the specific process of comparing the actual teaching content with the planned teaching content. This method can facilitate the current user to check for omissions and make up for the gaps in the completed courses, adjust and add the first teaching unit that was not taught in the first abnormal event to the next course or subsequent teaching plan information, realize comprehensive monitoring and management of knowledge points in the teaching courses, use the teaching evaluation text as a feedback file in the teaching plan information, realize dynamic adjustment of the teaching plan information, and improve the subsequent teaching quality.

[0195] In some embodiments, the method further comprises:

[0196] Perform text recognition on the current actual teaching content to determine whether there is keyword information, which includes at least one of social responsibility, moral concept, civic awareness, ability assessment, social response, and practical results;

[0197] If there is keyword information, the actual teaching content is input into the second neural network model to obtain a follow-up investigation plan corresponding to the actual teaching content, the follow-up investigation plan including at least one of the investigation method, investigation content, and tracking duration and tracking frequency;

[0198] Fill the follow-up survey plan into the table column of the second table tag associated with the follow-up survey plan, or add the follow-up survey plan as an attachment to the attachment information.

[0199] In this embodiment, the text recognition of the actual teaching content can be understood as the word segmentation of the text converted from the actual teaching content to obtain a number of words. The keyword information is used to identify a number of words to determine whether there are expressions such as social responsibility, moral concepts, civic awareness, ability assessment, social response, and practical results. It should be noted that the keyword information shows the expressions that need to be continuously tracked. If there is keyword information, it means that the current actual teaching content needs to be tracked after the course ends.

[0200] Furthermore, the actual teaching content is input into the second neural network model to obtain a follow-up survey plan that matches the current actual teaching content. The follow-up survey plan includes survey methods, inspection contents, tracking duration, and tracking frequency. Among them, the survey methods include assessment, questionnaires, social practice, etc., the inspection contents are inspection points generated based on the actual teaching content of the current course, the tracking duration is the duration of the current follow-up survey plan, and the tracking frequency is the number of repetitions of the follow-up survey plan.

[0201] Fill the follow-up survey plan into the table column of the second table tag associated with the follow-up survey plan, or add the follow-up survey plan as an attachment to the attachment information, so that users can use it according to actual needs.

[0202] This embodiment processes the actual teaching content through text processing and matches it with keyword information to determine whether the current actual teaching content requires subsequent follow-up investigation, and generates a follow-up investigation plan for user reference, thereby improving the efficiency of course follow-up after the course is completed.

[0203] See also Figure 5 In some embodiments, the planned teaching content, the actual teaching content, the teaching objectives, the actual teaching problems and the solutions are recorded as the text to be processed, and the planned teaching content, the actual teaching content, the teaching objectives, the actual teaching problems and the solutions are filled into the table column corresponding to the second table label according to the second typesetting specification information, and further includes:

[0204] S501, obtaining a second typesetting specification for each second table label in the second typesetting specification information, wherein the second typesetting specification information includes at least one of font size, word limit, line spacing, alignment, and font type;

[0205] S502, matching the text to be processed with the second table tags one by one, the text to be processed having a second initial typesetting format;

[0206] S503, determining one by one whether the current second initial typesetting format complies with the second typesetting specification;

[0207] S504: If yes, fill the text to be processed into the table column corresponding to the second table tag in the second initial typesetting format;

[0208] S505: if not, generate a second typesetting processing function model according to the second typesetting specification, and input the text to be processed into the second typesetting processing function model for format adjustment to obtain a first processed text;

[0209] S506, determining whether the number of words in the first processed text meets the second typesetting specification;

[0210] S507, if not, input the content of the first processed text into the third neural network model to obtain a second processed text with a limited number of words that meets the second typesetting specification;

[0211] S508, mapping and storing the first processed text and the second processed text, filling the second processed text into the table column corresponding to the second table label, and displaying the first processed text in the sidebar of the table column of the current second table label in the form of annotation;

[0212] S509: If yes, fill the first processed text into the table column corresponding to the second table label.

[0213] In step S501, each second table tag corresponds to a second typesetting specification, and in step S502, the text to be processed is matched with the second table tags one by one, and the text to be processed has a second initial typesetting format. It should be noted that the difference between the text to be processed and the text information value is that the text to be processed exists in the form of paragraphs or short sentences, which has an accurate semantic reference information, and the text information value is mostly a noun. The similarity between the text to be processed and the text information value is that the second initial typesetting format and the first initial typesetting format both refer to the default embodiment format of the text.

[0214] In step S503, determine whether the second initial typesetting format conforms to the second typesetting specification. In step S504, if the second initial typesetting format conforms to the second typesetting specification, directly fill the text to be processed in the second initial typesetting format into the table column corresponding to the second table label. In step S505, if the second initial typesetting format does not conform to the second typesetting specification, it is necessary to generate a second typesetting processing function model according to the second typesetting specification. Similar to the first typesetting processing function model, the second typesetting processing function model can adjust the format of the text to be processed to obtain the first processed text. It should be noted that the second typesetting processing function does not modify the number of words in the text to be processed.

[0215] Furthermore, in step S506, it is determined whether the word count of the first processed text meets the word count requirement in the second typesetting specification; if the word count of the first processed text does not meet the word count requirement in the second typesetting specification, the first processed text is input into the third neural network model, and the third neural network model reorganizes the words of the first processed text and retains the semantics of the first processed text to obtain a second processed text with a limited word count that meets the second typesetting specification.

[0216] In step S508, the first processed text and the second processed text are mapped and stored, and the second processed text is filled in the table column corresponding to the second table tag. At the same time, the first processed text is displayed in the sidebar of the table column of the current second table tag in the form of annotations for user reference.

[0217] In step S509, if the number of words in the first processed text meets the limited number of words in the second typesetting specification, the first processed text is directly filled into the table column corresponding to the second table tag.

[0218] Furthermore, in some embodiments, the method further comprises:

[0219] Determine whether the number of annotations in the current second table document exceeds a preset annotation number threshold;

[0220] If yes, the second table labels containing the annotations are sorted and an annotation index is generated;

[0221] Displaying the annotation index in a sidebar of the second table document;

[0222] Obtain the input command in the sidebar and identify the input coordinates of the input command;

[0223] Match the second table tag within the current annotation index according to the typed coordinates;

[0224] The first processing text associated with the second table tag is displayed in a pop-up window.

[0225] This embodiment sets a preset threshold for the number of annotations to avoid the problem of multiple first processing texts overlapping in the sidebar when there are too many annotations, and generates an annotation index based on the annotations, which is displayed in the sidebar in the form of an annotation index. When the user uses a mouse, keyboard, touch screen or other input device to input a typing instruction, the second table label selected in the current annotation index is identified based on the typing of the typing instruction, and the first processing text associated with the second table label is displayed in a pop-up window. This method makes it easy for users to check the content corresponding to the second table label after the teaching evaluation text is generated. The setting of the annotation index also makes the entire inspection process more convenient and improves the user experience comfort.

[0226] In some embodiments, the first neural network model is trained in the following manner:

[0227] Constructing a first convolutional neural network model, and extracting historical teaching problems and historical solutions from a database to generate a first sample database, wherein first sample data in the first sample database includes sample teaching problems and sample solutions;

[0228] Selecting first sample data to train a first convolutional neural network model until the accuracy of the test solution output by the first convolutional neural network model and the sample solution reaches a first preset threshold, thereby obtaining a first neural network model;

[0229] The second neural network model is trained in the following way:

[0230] Constructing a second convolutional neural network model, and extracting historical actual teaching content and historical tracking survey plans from the database to generate a second sample database, wherein the second sample data in the second sample database includes sample actual teaching content and sample tracking survey plans;

[0231] Selecting second sample data to train the second convolutional neural network model until the accuracy of the test tracking survey plan and the sample tracking survey plan output by the second convolutional neural network model reaches a second preset threshold, thereby obtaining a second neural network model;

[0232] The third neural network model is trained in the following way:

[0233] Constructing a third convolutional neural network model, and extracting historical processing text and historical summary text from the database to generate a third sample database, wherein the third sample data in the third sample database includes the sample processing text and the sample summary text;

[0234] A third sample data is selected to train the third convolutional neural network model until the accuracy of the test summary text and the sample summary text output by the third convolutional neural network model reaches a third preset threshold, thereby obtaining a third neural network model.

[0235] This embodiment obtains basic information, which includes teacher information, course information, project information and declaration information, and selects a preset template in the database according to the declaration information, generates a first screening condition according to the first table tag group, and performs data screening on the basic information to obtain a first basic text, and fills the first basic text into the table column corresponding to the first table tag according to the first typesetting specification information to obtain a first table document; generates a second screening condition according to the second table tag group, and performs data screening on the basic information to obtain a second basic text, the second basic text includes planned teaching content, actual teaching content and teaching objectives; compares the differences between the planned teaching content and the actual teaching content to obtain the current course. actual teaching problems; inputting the actual teaching problems into the first neural network model to obtain a solution plan corresponding to the current actual teaching problems, the solution plan including at least one of a solution suggestion and a solution to improve or solve the current actual teaching problems; filling the planned teaching content, actual teaching content, teaching objectives, actual teaching problems and the solution plan into the table column corresponding to the second table label according to the second typesetting specification information to obtain a second table document; generating a final application form based on the first table document and the second table document; exporting the basic information to a file according to the data requirement information and generating attachment information, the attachment information including an attachment list and attached files; mapping and storing the final application form and the attachment information to obtain a teaching evaluation text. Teachers can summarize the required file materials into basic information in the form of folders or documents, and then execute the corresponding logical steps to selectively screen and typeset the contents of the basic information to obtain the teaching evaluation text. In this process, the teaching content in the basic information is carefully semantically analyzed and evaluated to obtain the second basic text, which is then filled in the second table document corresponding to the second table label. This method can greatly reduce the workload of teachers' independent editing of documents during the writing process of teaching evaluation documents, save editing time, and significantly improve the writing efficiency and quality of teaching evaluation texts.

[0236] Finally, it should be noted that although the above embodiments have been described in the specification and drawings of the present invention, the patent protection scope of the present invention cannot be limited thereby. All technical solutions generated by replacing or modifying equivalent structures or equivalent processes based on the essential concept of the present invention and using the contents recorded in the specification and drawings of the present invention, as well as directly or indirectly implementing the technical solutions of the above embodiments in other related technical fields, are included in the patent protection scope of the present invention.

Claims

1. A method for automatically generating teaching evaluation text, characterized in that: include: Obtaining basic information, wherein the basic information includes teacher information, course information, project information and application information. The teacher information includes teacher name, teacher number, education information, teaching experience information, teacher rating information, teaching achievement information and teaching honor information. The course information includes subject information, planned teaching information of the current course and actual teaching information of the current course. The project information includes the teaching plan information to which the current course belongs. The application information includes the application file name and the application file number. Select a preset template in the database according to the declaration information, the preset template includes at least one of declaration form information, typesetting specification information, and data requirement information, the declaration form information includes a form tag and a form column, each of the form tags corresponds to a form column, the form tags include a first form tag group and a second form tag group, the first form tag group includes a plurality of first form tags, the second form tag group includes a plurality of second form tags, and the typesetting specification information includes first typesetting specification information and second typesetting specification information; Generate a first screening condition according to the first table tag group, perform data screening on the basic information to obtain a first basic text, and fill the first basic text into the table column corresponding to the first table tag according to the first typesetting specification information to obtain a first table document; Generate a second screening condition according to the second table tag group, and perform data screening on the basic information to obtain a second basic text, wherein the second basic text includes planned teaching content, actual teaching content, and teaching objectives; perform a difference comparison between the planned teaching content and the actual teaching content to obtain actual teaching problems of the current course; Inputting the actual teaching problem into the first neural network model to obtain a solution plan corresponding to the current actual teaching problem, wherein the solution plan includes at least one of a solution suggestion and a solution method for improving or solving the current actual teaching problem; Fill the planned teaching content, actual teaching content, teaching objectives, actual teaching problems and solutions into the table column corresponding to the second table label according to the second typesetting specification information to obtain a second table document; Generate a final declaration form based on the first form document and the second form document; Export the basic information to a file according to the data requirement information and generate attachment information, wherein the attachment information includes an attachment list and attached files; Mapping and storing the final declaration form and the attachment information to obtain a teaching evaluation text; Generating the first screening condition according to the first table tag group includes: Acquire the meaning of the first table label, and obtain the data type corresponding to the first table label according to the meaning of the first table label, wherein the data type includes at least one of numbers, words, letters, and punctuation marks; Mapping and storing the meaning of the first table label and the data type to obtain a current screening index of the first table label; Arrange the screening indicators of the plurality of the first table tags according to the arrangement order in the first table tag group to obtain the first screening condition; The basic information is screened to obtain the first basic text including: Converting the first screening condition into an SQL screening sentence segment, and running an application program containing the SQL screening sentence segment on the basic information to obtain a text information value that meets the first screening condition, the text information value having a first initial typesetting format, and arranging and merging the text information value to obtain the first basic text; Filling the first basic text into the table column corresponding to the first table tag according to the first typesetting specification information to obtain the first table document includes: Obtaining a first typesetting specification for each of the first table tags in the first typesetting specification information, wherein the first typesetting specification information includes at least one of font size, word limit, data type, line spacing, alignment, and font type; Matching the text information values ​​in the first basic text with the first table tags one by one, and determining whether the current first initial typesetting format complies with the first typesetting specification; If yes, then fill the text information value into the table column corresponding to the first table tag in the first initial typesetting format; If not, generating a first typesetting processing function model according to the first typesetting specification, inputting the text information value into the first typesetting processing function model for format adjustment, and filling the format-adjusted text information value into the table column corresponding to the first table tag; The second filtering condition is generated according to the second table tag group, and the basic information is filtered to obtain the second basic text including: The meaning of the second table label is obtained and recorded as a screening text topic, and a plurality of the screening text topics are sorted according to the arrangement order in the second table label group to obtain the second screening condition.

2. The method for automatically generating teaching evaluation text according to claim 1, characterized in that: The second filtering condition is generated according to the second table tag group, and the basic information is filtered to obtain the second basic text including: Preprocessing the basic information to obtain basic information containing only paragraphs, which is recorded as a first intermediate text, and the paragraph in the first intermediate text is recorded as a first paragraph; Inputting the first paragraphs into the semantic analysis model one by one to obtain the paragraph text topics of the first paragraphs, and mapping and storing the paragraph text topics with the first paragraphs; Performing topic clustering on the multiple paragraph text topics to obtain a number of paragraph topic sets, each of which has a central topic; The central theme is matched with the screening text theme in the second screening condition in terms of similarity, and the first paragraphs associated with the central theme having the same similarity with the screening text theme are sorted and merged to obtain the second basic text.

3. The method for automatically generating teaching evaluation text according to claim 2, characterized in that: The multiple paragraph text topics are clustered to obtain several paragraph topic sets, each of which has a central topic including: Set K first cluster center points, and calculate the first distance value between the paragraph text topic and each of the first cluster center points one by one; According to the first distance value, the paragraph text topic is assigned to the cluster where the first cluster center point closest to the paragraph text topic is located, which is recorded as the first cluster cluster; Calculate the average vector of all paragraph text topics in each of the first clusters as the second cluster center point of the current first cluster, calculate the second distance value between the paragraph text topic and each of the second cluster center points one by one, and assign the paragraph text topic to the cluster where the second cluster center point closest to it is located according to the second distance value, which is recorded as the second cluster cluster; Repeat the above steps until a preset number of iterations is reached to obtain a plurality of final clusters, wherein the paragraph text topics in each of the final clusters constitute a paragraph topic set; Inputting the paragraph text topics in the paragraph topic set into the semantic analysis model to obtain the central topic of the current paragraph topic set; The central theme is matched with the screening text theme in the second screening condition by similarity, and the first paragraphs associated with the central theme having the same similarity with the screening text theme are sorted and merged to obtain the second basic text, which further includes: Converting the central topic into a central topic vector, and converting the screening text topic into a screening text topic vector; Calculate the cosine similarity algorithm between the selected text topic vector and the central topic vector to obtain the cosine similarity between each central topic vector and the selected text topic vector; Obtaining a similarity threshold corresponding to the current screening text topic, determining whether the cosine similarity between each of the central topics and the current screening text topic is within the range of the similarity threshold, and if so, arranging and merging the first paragraph associated with the central topic to obtain text information of the current screening text topic, recorded as screening text information; Repeat the above steps until all the filter text topics are traversed to obtain multiple filter text information; Arrange and merge the plurality of the filtered text information to obtain the second basic text.

4. The method for automatically generating teaching evaluation text according to claim 3, characterized in that: Compare the planned teaching content and the actual teaching content to obtain the actual teaching problems of the current course, including: The planned teaching content is divided into the smallest teaching units and arranged in the planned teaching order to obtain a first teaching timeline, wherein the first teaching timeline is provided with a plurality of first teaching units arranged in sequence; the actual teaching content is divided into the smallest teaching units and arranged in the actual teaching order to obtain a second teaching timeline, wherein the second teaching timeline is provided with a plurality of second teaching units arranged in sequence; Matching the first teaching timeline with the second teaching timeline one by one in chronological order to obtain a plurality of first abnormal events, wherein the first abnormal event includes an abnormal time node and an abnormal event, and the abnormal event includes at least one of a first teaching unit that has not been taught and a second teaching unit that has been newly taught; Predicting the causes of the first abnormal events one by one to obtain a plurality of first abnormal causes, wherein the first abnormal cause includes at least one of a student factor, a teacher factor, a classroom factor, and a time factor; The first abnormal cause is sorted out to obtain the actual teaching problem.

5. The method for automatically generating teaching evaluation text according to claim 4, characterized in that: The method further comprises: Obtain the planned teaching information in the course information of the next course in the current teaching plan information, and record it as pending teaching information; In the solution plan, the teaching units that need to be taught again in the course information of the next course are screened out and recorded as abnormal teaching units; Determining whether the number of the abnormal teaching units exceeds the upper limit of the number of newly added teaching units; If yes, then adding course information to the teaching plan information, and generating new planned teaching information of the new course information according to the abnormal teaching unit; If not, the abnormal teaching unit is inserted into the pending teaching information and updated to become the scheduled teaching information in the next course information.

6. The method for automatically generating teaching evaluation text according to claim 5, characterized in that: The method further comprises: Performing text recognition on the current actual teaching content to determine whether there is keyword information, wherein the keyword information includes at least one of social responsibility, moral concept, civic awareness, ability assessment, social response, and practical results; If there is keyword information, the actual teaching content is input into the second neural network model to obtain a follow-up investigation plan corresponding to the actual teaching content, wherein the follow-up investigation plan includes at least one of the investigation method, investigation content, and tracking duration and tracking frequency; The follow-up survey plan is filled into a table column of a second table tag associated with the follow-up survey plan, or the follow-up survey plan is added as an attachment to the attachment information.

7. The method for automatically generating teaching evaluation text according to claim 6, characterized in that: The planned teaching content, actual teaching content, teaching objectives, actual teaching problems and solutions are recorded as text to be processed, and the planned teaching content, actual teaching content, teaching objectives, actual teaching problems and solutions are filled into the table column corresponding to the second table label according to the second typesetting specification information, and the following is also included: Obtaining the second typesetting specification for each of the second table tags in the second typesetting specification information, wherein the second typesetting specification information includes at least one of font size, word limit, line spacing, alignment, and font type; Matching the text to be processed with the second table tags one by one, wherein the text to be processed has a second initial typesetting format; Determining whether the second initial typesetting format currently meets the second typesetting specification one by one; If yes, then fill the text to be processed into the table column corresponding to the second table tag in the second initial typesetting format; If not, generating a second typesetting processing function model according to the second typesetting specification, and inputting the text to be processed into the second typesetting processing function model for format adjustment to obtain a first processed text; Determine whether the word count of the first processed text meets the second typesetting specification, and if not, input the content of the first processed text into a third neural network model to obtain a second processed text with a limited word count that meets the second typesetting specification; The first processed text and the second processed text are mapped and stored, and the second processed text is filled into the table column corresponding to the second table label, and the first processed text is displayed in the sidebar of the table column of the current second table label in the form of annotations.

8. The method for automatically generating teaching evaluation text according to claim 7, characterized in that: The method further comprises: Determining whether the number of annotations in the second table document currently exceeds a preset annotation number threshold; If yes, then sorting the second table labels containing the annotations and generating an annotation index; Displaying the annotation index in a sidebar of the second table document; Acquire a typing instruction in the sidebar, and identify a typing coordinate of the typing instruction; Matching a second table tag in the current annotation index according to the input coordinates; The first processed text associated with the second table tag is displayed in a pop-up window.

9. The method for automatically generating teaching evaluation text according to claim 8, characterized in that: The first neural network model is trained by the following method: Constructing a first convolutional neural network model, and extracting historical teaching problems and historical solutions from the database to generate a first sample database, wherein the first sample data in the first sample database includes sample teaching problems and sample solutions; Selecting the first sample data to train the first convolutional neural network model until the accuracy of the test solution output by the first convolutional neural network model and the sample solution reaches a first preset threshold, thereby obtaining the first neural network model; The second neural network model is trained by the following method: Constructing a second convolutional neural network model, and extracting historical actual teaching content and historical tracking survey plans from the database to generate a second sample database, wherein the second sample data in the second sample database includes sample actual teaching content and sample tracking survey plans; Selecting the second sample data to train the second convolutional neural network model until the accuracy of the test tracking survey plan output by the second convolutional neural network model and the sample tracking survey plan reaches a second preset threshold, thereby obtaining the second neural network model; The third neural network model is trained by the following method: Constructing a third convolutional neural network model, and extracting historical processing text and historical summary text from the database to generate a third sample database, wherein the third sample data in the third sample database includes sample processing text and sample summary text; The third sample data is selected to train the third convolutional neural network model until the accuracy of the test summary text output by the third convolutional neural network model and the sample summary text reaches a third preset threshold, thereby obtaining the third neural network model.

Images