A label matching method and system for smart teaching

Abstract

This application relates to the technical field of online education technology, and provides a label matching method and system for smart teaching, the method comprising: marking the exercises with original labels based on each knowledge point and the exercises matched by the knowledge point; during the exercise practice process, obtaining custom labels, and custom-marking the exercises based on the custom labels, the custom labels including custom semantic labels and custom keyword labels; during the exercise retrieval process, obtaining an exercise retrieval formula, the exercise retrieval formula including at least standard search terms corresponding to the original labels and custom search terms corresponding to the custom labels; and generating an exercise recommendation scheme based on the exercise retrieval formula, custom labels, original labels, and student historical exercise records. The method can improve the matching degree between students and exercises and increase the student's retrieval hit rate for exercises.

Description

Technical Field

[0001] The present application relates to the field of online education technology, and in particular to a label matching method and system for smart teaching. Background Art

[0002] Online education is e-Learning, also known as distance education or online learning. In the current concept, it generally refers to a network-based learning behavior, which is similar to the concept of online training.

[0003] With the rapid development of computer technology, online education has expanded in multiple directions. Students are particularly drawn to the ability to find similar exercises based on a single exercise. For example, if a student struggles with a certain type of exercise or a specific test point, they need to find more similar exercises to consolidate and strengthen their understanding. Finding similar exercises to a specific exercise requires extracting test point information from the exercise. Test point extraction refers to the process of discovering, identifying, and ultimately extracting the concepts, key points, and rules inherent in the exercise information, thereby changing the traditional organization and management of knowledge points and test points.

[0004] However, current selection of similar exercises is generally based on data retrieval based on the fixed exercise tags in the teaching platform. For example, a student enters a known knowledge point, and the teaching platform then uses the knowledge point to provide the student with exercises that match the knowledge point. Another example is a student entering a text description, and the teaching platform extracts keywords that match the knowledge point from the text description, then uses the keyword search to retrieve the corresponding exercises and provide them to the student. With this fixed exercise retrieval model, the convenience of exercise retrieval depends on whether the teaching platform's tag system for matching exercises can cover all knowledge points and whether students can easily use the tag system to search. This greatly limits the retrieval of exercises and reduces the practical effectiveness of the exercise retrieval function. Summary of the Invention

[0005] The purpose of this application is to provide a label matching method and system for smart teaching, which can improve the matching degree between students and exercises and increase the students' retrieval hit rate of exercises.

[0006] The above-mentioned application objectives of this application are achieved through the following technical solutions:

[0007] In a first aspect, the present application provides a label matching method for smart teaching using the following technical solution.

[0008] A label matching method for smart teaching, the method comprising:

[0009] Marking the exercises with original labels according to the knowledge points and the exercises that match the knowledge points;

[0010] During the exercise practice, a custom tag is obtained, and the exercise is custom-marked according to the custom tag, wherein the custom tag includes a custom semantic tag and a custom keyword tag;

[0011] During the exercise search process, an exercise search formula is obtained, wherein the exercise search formula includes at least a standard search term corresponding to the original label and a custom search term corresponding to the custom label;

[0012] Generate an exercise recommendation plan based on the exercise search formula, the custom tag original tag and the student's historical exercise record.

[0013] In a preferred embodiment, the method for obtaining a custom tag includes:

[0014] Obtaining custom options, wherein the custom options include semantic options and keyword options;

[0015] The custom label is generated according to the custom options and the custom entry information.

[0016] In a preferred embodiment, the method for generating a custom label based on the custom options and custom input information includes:

[0017] When the custom option is a semantic option, identifying multiple semantic keywords in the custom input information, and generating a custom semantic tag based on the current exercise content and the multiple semantic keywords;

[0018] When the custom option is a keyword option, the custom keyword in the custom entry information is used as the custom keyword tag.

[0019] In a preferred embodiment, the method of identifying multiple semantic keywords in the custom input information and generating a custom semantic tag based on the current exercise content and the multiple semantic keywords includes:

[0020] Identifying standard words in the custom search terms based on the current exercise content;

[0021] Obtaining the plurality of semantic keywords from the selected standard words and the entered supplementary words;

[0022] The custom semantic tag is obtained by logically arranging multiple semantic keywords.

[0023] In a preferred embodiment, the selected standard words are obtained by students selecting from the standard words.

[0024] In a preferred embodiment, the self-defined semantic tag is obtained by logically arranging multiple semantic keywords, including:

[0025] Students logically sort multiple semantic keywords and add logical conjunctions between the semantic keywords to obtain the custom semantic labels.

[0026] In a preferred embodiment, the method further comprises:

[0027] Storing the custom semantic tags and custom keyword tags in a semantic tag set and a keyword tag set respectively;

[0028] When obtaining an exercise search formula, matching custom tags are retrieved in a semantic tag set or a keyword tag set based on the custom search term and the selected custom classification. The custom classification includes a semantic class corresponding to the semantic tag set and a keyword class corresponding to the keyword tag set.

[0029] In a preferred embodiment, the method of custom-marking the exercises according to the custom tags includes:

[0030] When the custom tag is a custom keyword tag, identifying the logical position of the custom keyword in the current exercise, and marking the exercise containing the custom keyword and having the same logical position of the custom keyword in the exercise with the custom keyword as the custom keyword tag;

[0031] When the custom tag is a custom semantic tag, the exercises that have the multiple semantic keywords and each semantic keyword has the same logical position as the semantic keyword in the current exercise are marked with the custom semantic tag.

[0032] In a preferred embodiment, generating an exercise recommendation scheme based on the exercise search formula, the custom tag, the original tag, and the student's historical exercise records includes:

[0033] Obtaining standard search terms and custom search terms in the exercise search formula;

[0034] Locating the first search data set based on the standard search term;

[0035] Searching the first search data set for exercises with the custom label according to the custom search term to obtain a second search data set;

[0036] Obtaining related exercises in the second retrieval data set according to the student's historical exercise records;

[0037] Generate exercise recommendation solutions based on related exercises and the second search dataset.

[0038] On the second aspect, the label matching system for smart teaching provided in this application adopts the following technical solution.

[0039] A label matching system for smart teaching, the system comprising:

[0040] An original label storage module is used to store original marking labels for exercises, wherein the original marking labels are labels for the exercises based on the knowledge points and the exercises matched with the knowledge points;

[0041] A custom tag marking module is used to obtain custom tags during the exercise process and perform custom tags on the exercises according to the custom tags, wherein the custom tags include custom semantic tags and custom keyword tags;

[0042] A retrieval module is used to obtain a question retrieval formula during the question retrieval process, wherein the question retrieval formula includes at least a standard retrieval term corresponding to the original label and a custom retrieval term corresponding to the custom label;

[0043] The solution recommendation module is used to generate exercise recommendation solutions based on the exercise search formula, custom tags and original tags.

[0044] To sum up, this application includes the following beneficial effects: the setting of original tags can realize the basic tag configuration of exercises. During the daily learning process of students, students can label exercises independently, that is, enter custom tags, and customize exercises through custom tags. Therefore, during the learning process of students, similar exercises can be retrieved according to the combination of standard search terms and custom search terms. Compared with searching only by knowledge points, the retrieval hit rate of students for similar exercises is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a flowchart of the label matching method for smart teaching in an embodiment of the present application.

[0046] Figure 2 This is a diagram of the label matching system for smart teaching in an embodiment of the present application. DETAILED DESCRIPTION

[0047] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0048] In this document, the term "and / or" simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " in this document, unless otherwise specified, generally indicates an "or" relationship between the related objects.

[0049] This application provides a tag matching method for smart teaching, which is applied to exercise tag matching on online teaching platforms, especially in the field of mathematics. The online teaching platform includes but is not limited to smart terminals such as mobile phones, computers, and iPads.

[0050] During the online education process, students can use the smart terminal to at least connect with exercises, enter exercise tags, and search for exercises. For example, during the connection process, students can enter tags that match the current exercise so that they can later search for the current exercise by tags.

[0051] Reference Figure 1 , the tag matching method comprises the following steps:

[0052] Step S100: marking the exercises with original labels according to the knowledge points and the exercises matched to the knowledge points;

[0053] Step S200: During the exercise practice process, obtain custom tags, and perform custom marking on the exercise according to the custom tags, wherein the custom tags include custom semantic tags and custom keyword tags;

[0054] Step S300: During the exercise search process, an exercise search formula is obtained, wherein the exercise search formula includes at least a standard search term corresponding to the original tag and a custom search term corresponding to the custom tag;

[0055] Step S400: Generate exercise recommendation solutions based on the exercise search formula, custom tags and original tags.

[0056] In step S100, the original label is the knowledge point. During the construction of the teaching platform, exercises need to be entered into the teaching platform. When entering exercises, it is necessary to select the knowledge point that the exercises can be classified into. There are one or more knowledge points that an exercise can be classified into, so the original label corresponding to an exercise also has one or more. In a preferred example, the knowledge point is the same as the course knowledge point in the education system. Students can select the corresponding knowledge point system by choosing the textbooks used for learning, and then match the knowledge points corresponding to the textbooks.

[0057] In step S200, during the process of practicing exercises, if the student is not proficient in the knowledge points of the current exercise, the student will mark the current exercise by entering a custom label. Among them, in the online teaching platform, the interactive interface of the exercise practice is configured with a label entry bar, and the student marks the current exercise through the label entry bar. The label entry bar is configured with custom options, and the custom options include semantic options and keyword options. That is, when the student clicks on the label entry bar, the online teaching platform will feedback the selection prompts "custom keywords" and "custom semantics". When the student selects the custom keyword, the student enters a fixed keyword through the label entry bar. When the student selects the custom semantics, the student enters a piece of semantic text through the label entry bar. It should be noted that the present application does not make a unique limitation on the selection prompt method of the online teaching platform feedback, as long as it can enable students to select the type of custom label to be entered. For example, the label entry bar can also include a keyword label entry area that matches the custom keyword and a semantic label entry area that matches the custom semantics.

[0058] In this application scheme, the information entered by students in the label input column is recorded by the teaching platform as custom input information.

[0059] When the custom option is a keyword option, the custom keyword in the custom input information is used as the custom keyword tag. Although the custom keyword is entered during the current exercise, after obtaining the custom keyword tag, the teaching platform marks the exercises in the exercise library according to the custom keyword tag. The method of marking exercises according to the custom keyword tag is as follows:

[0060] The teaching platform is equipped with a logical analysis model, which analyzes the logical relationships of each exercise using the model. The logical relationship categories include total-to-specific, primary-secondary, parallel, progressive, point-to-surface, cause-and-effect, and qualitative and quantitative. After determining the logical relationship category of the exercise, the exercise is divided according to its logical position. For example, in Exercise X, "Given points A (1, 2) and B (3, 1), the equation of the perpendicular bisector of line segment AB is ()" is classified as qualitative, meaning that a fixed result is derived from a known premise. The logical positions are respectively the known premise and the fixed result. The content of the known premise portion is "Given points A (1, 2) and B (3, 1)", and the content of the fixed result portion is "The equation of the perpendicular bisector of line segment AB."

[0061] The logical analysis model is a semantic segmentation model. Existing exercises are used as training samples. After the training samples are classified according to logical relationship categories, the content of the samples is divided into logical positions, and the samples are entered into the semantic segmentation model for training to obtain the logical analysis model.

[0062] After obtaining the custom keyword tag, by analyzing the logical position of the custom keyword in the current exercise, exercises containing the custom keyword and having the same logical position of the custom keyword in the exercise are retrieved, and the retrieved exercises are marked with the custom keyword tag.

[0063] For example, the current exercise is "Given points M (5, 7) and N (3, 8), find the equation of the parabola () passing through the origin, M, and N." The entered custom keyword is "Known Points." If the position of the custom keyword in the current exercise is the same as its logical position in Exercise X, Exercise X will be labeled with the "Known Points" custom keyword tag.

[0064] In a preferred example, the custom keywords entered during the practice of the current exercise are not applicable. The interactive interface of the teaching platform is configured with a custom keyword deletion function, and the currently entered custom keywords can be deleted through the custom keyword deletion function.

[0065] When the custom option is a semantic option, multiple semantic keywords in the custom input information are identified, and a custom semantic tag is generated based on the current exercise content and the multiple semantic keywords. The method for generating the custom semantic tag is as follows:

[0066] wherein the standard words in the custom search terms are identified based on the current exercise content;

[0067] Obtaining the plurality of semantic keywords from the selected standard words and the entered supplementary words;

[0068] The custom semantic tag is obtained by logically arranging multiple semantic keywords.

[0069] The entered custom search terms include standard terms that match the field of the exercise. In the embodiment of the present application, the standard terms are standard terms in the field of mathematics. After entering the custom search terms, the standard terms are first identified and displayed to the student in a floating window. The student selects the required standard terms from the standard terms through the floating window and enters the required supplementary terms to obtain multiple semantic keywords. For example, the current exercise is "Given points M (5, 7) and N (3, 8), find the equation of the parabola () passing through the origin, M, and N." The custom search term entered by the student is "Calculate the equation of the parabola from known points." The standard terms obtained through the custom search terms include: "Given points M (5, 7) and N (3, 8)", and "Equation". The student can add the supplementary term "two" in the floating window to form multiple semantic keywords: "two", "Given points M (5, 7) and N (3, 8)", and "Equation".

[0070] When arranging the logical relationships of multiple semantic keywords, students logically sort the multiple semantic keywords and add logical conjunctions between the semantic keywords to obtain the custom semantic labels. The logical conjunctions include but are not limited to: and, or, not, equal, greater than, less than.

[0071] Furthermore, the floating window contains the logical relationship category and logical position of the current exercise. When students logically sort multiple semantic keywords, the corresponding semantic keywords are placed in the matching logical positions and then the semantic keywords in the corresponding logical positions are logically sorted. That is, the custom semantic label contains the semantic keywords and logically associated words.

[0072] After obtaining the custom semantic tags and custom keyword tags, the teaching platform first stores the custom semantic tags and custom keyword tags in the data storage area corresponding to the student, and stores the custom semantic tags and custom keyword tags in the semantic tag set and the keyword tag set respectively.

[0073] When the custom tag is a custom semantic tag, the exercises that have the multiple semantic keywords and each semantic keyword has the same logical position as the semantic keyword in the current exercise are marked with the custom semantic tag.

[0074] When obtaining an exercise search formula, matching custom tags are retrieved in a semantic tag set or a keyword tag set based on the custom search term and the selected custom classification. The custom classification includes a semantic class corresponding to the semantic tag set and a keyword class corresponding to the keyword tag set.

[0075] It should be noted that when the custom classification is semantic, the custom search term entered by the student includes the custom semantic tag, but is not limited to being the same as the custom semantic tag. For example, if the custom search terms include "two" and "known point", all custom semantic tags containing "two" and "known point" will be fed back to the teaching platform's interactive interface in a list format, and the student will then select a custom semantic tag from the list as the final custom search term.

[0076] Furthermore, the exercise retrieval formula includes standard search terms. When obtaining the exercise retrieval formula, the first retrieval data set corresponding to the standard search terms is first locked through the standard search terms. In the first retrieval data set, the matching custom tags are retrieved in the semantic tag set or keyword tag set based on the custom search terms and the selected custom classification to obtain the second retrieval data set. The related exercises in the second retrieval data set are obtained based on the student's historical exercise records; and exercise recommendation plans are generated based on the related exercises and the second retrieval data set.

[0077] In a preferred example, in the exercise recommendation scheme, the logical associations between the semantic keywords are explicitly marked, and the students click on the corresponding logical associations. The teaching platform then performs semantic analysis on the exercises in the second retrieval data set based on the selected logical associations. If the logical relationship between the two semantic keywords connected by the logical associations does not match the logical associations, the exercises will be deleted from the second retrieval data set.

[0078] Furthermore, when the custom classification is selected as the semantic category, the list fed back by the interactive interface of the teaching platform includes the logical relationship category, and students filter out the custom semantic tags with the corresponding logical relationship category in the list through the logical relationship category option.

[0079] During the student's practice, the teaching platform records the student's past practice information and marks the questions the student has completed as correct / incorrect. The associated exercises are the exercises in the second search data set that contain the correct / incorrect mark. When generating the recommended exercise plan, the interactive interface is configured with a correct / incorrect option. Based on the selected "correct" option and "incorrect" option, the interactive interface sorts the associated exercises marked "correct" or "incorrect" at the top of the list to generate the recommended exercise plan.

[0080] By adopting the above technical solution: the setting of original tags can realize the basic tag configuration of exercises. In the daily learning process of students, students can label exercises independently, that is, enter custom tags, and customize exercises with custom tags. In this way, students can search for similar exercises based on the combination of standard search terms and custom search terms during the learning process. Compared with searching only by knowledge points, the search hit rate of students for similar exercises is improved.

[0081] Reference Figure 2 In another preferred embodiment, the present application further provides a label matching system for smart teaching, the system comprising:

[0082] An original label storage module is used to store original marking labels for exercises, wherein the original marking labels are labels for the exercises based on the knowledge points and the exercises matched with the knowledge points;

[0083] A custom tag marking module is used to obtain custom tags during the exercise process and perform custom tags on the exercises according to the custom tags, wherein the custom tags include custom semantic tags and custom keyword tags;

[0084] A retrieval module is used to obtain a question retrieval formula during the question retrieval process, wherein the question retrieval formula includes at least a standard retrieval term corresponding to the original label and a custom retrieval term corresponding to the custom label;

[0085] The solution recommendation module is used to generate exercise recommendation solutions based on the exercise search formula, custom tags and original tags.

[0086] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working process of the system module described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0087] In the several embodiments provided in this application, it should be understood that the disclosed system modules can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the module division is merely a logical functional division. In actual implementation, other division methods may be used, such as combining or integrating multiple modules or components into another system, or omitting or not implementing certain features.

[0088] The modules described as separate components may or may not be physically separated, and the components as functional modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple physical structures, especially the original label storage module and the custom label marking module. The original label stored in the original label storage module is applicable to all students on the teaching platform, but the custom label stored in the custom label marking module is only applicable to the corresponding students. Therefore, the original label storage module needs to be arranged in the integrated memory of the platform, while the custom label marking module can be arranged in the terminal used by the students without knowing the integrated processor.

[0089] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

Claims

1. A label matching method for smart teaching, characterized in that: The method comprises: Marking the exercises with original labels according to the knowledge points and the exercises that match the knowledge points; During the exercise practice, a custom tag is obtained, and the exercise is custom-marked according to the custom tag, wherein the custom tag includes a custom semantic tag and a custom keyword tag; During the exercise search process, an exercise search formula is obtained, wherein the exercise search formula includes at least a standard search term corresponding to the original label and a custom search term corresponding to the custom label; Generate an exercise recommendation plan based on the exercise search formula, the custom tag, the original tag and the student's historical exercise record; Methods for obtaining custom labels include: Obtaining custom options, wherein the custom options include semantic options and keyword options; Generating the custom label according to the custom options and custom entry information; The method for generating a custom label based on the custom options and custom input information includes: When the custom option is a semantic option, identifying multiple semantic keywords in the custom input information, and generating a custom semantic tag based on the current exercise content and the multiple semantic keywords; When the custom option is a keyword option, the custom keyword in the custom entry information is used as the custom keyword label; The method of identifying multiple semantic keywords in the custom input information and generating a custom semantic tag based on the current exercise content and the multiple semantic keywords includes: Identifying standard words in the custom search terms based on the current exercise content; Obtaining the plurality of semantic keywords from the selected standard words and the entered supplementary words; Arranging the multiple semantic keywords into logical relationships to obtain the custom semantic tag; The selected standard words are obtained by students choosing from the standard words; Arranging the logical relationships of multiple semantic keywords to obtain the custom semantic tag includes: The students logically sort the multiple semantic keywords and add logical conjunctions between the semantic keywords to obtain the custom semantic labels; The method further comprises: Storing the custom semantic tags and custom keyword tags in a semantic tag set and a keyword tag set respectively; When obtaining a question search formula, searching for matching custom tags in a semantic tag set or a keyword tag set based on the custom search term and the selected custom classification, wherein the custom classification includes a semantic class corresponding to the semantic tag set and a keyword class corresponding to the keyword tag set; The method for custom-marking the exercises according to the custom tags includes performing a logical relationship analysis on each exercise using a logical analysis model, analyzing the logical position of the custom keyword in the current exercise, searching for exercises containing the custom keyword and having the same logical position of the custom keyword in the exercise, and marking the retrieved exercises with the custom keyword tag: When the custom tag is a custom keyword tag, identifying the logical position of the custom keyword in the current exercise, and marking the exercise containing the custom keyword and having the same logical position of the custom keyword in the exercise with the custom keyword as the custom keyword tag; When the custom tag is a custom semantic tag, the exercises that simultaneously have the multiple semantic keywords and each of which has the same logical position as the semantic keyword in the current exercise are marked with the custom semantic tag; The logical analysis model is a semantic segmentation model. Existing exercises are used as training samples. After the training samples are classified according to logical relationship categories, the content of the samples is divided into logical positions. The samples are input into the semantic segmentation model for training to obtain the logical analysis model. The logical analysis model is used to analyze the logical relationships of each exercise. The logical relationship categories include total-to-specific, primary-secondary, parallel, progressive, point-to-surface, cause-and-effect, qualitative and quantitative. After the logical relationship categories of the exercises are determined, the exercises are divided according to their logical positions. When entering custom search terms, the standard terms are displayed to students in a floating window, which contains the logical relationship category and logical position of the current exercise. When students logically sort multiple semantic keywords, the corresponding semantic keywords are placed in the matching logical positions and then the semantic keywords in the corresponding logical positions are logically sorted.

2. The label matching method for smart teaching according to claim 1, characterized in that: Generating an exercise recommendation scheme based on the exercise search formula, custom tags, original tags, and student historical exercise records includes: Obtaining standard search terms and custom search terms in the exercise search formula; Locating the first search data set based on the standard search term; Searching the first search data set for exercises with the custom label according to the custom search term to obtain a second search data set; Obtaining related exercises in the second retrieval data set according to the student's historical exercise records; Generate exercise recommendation solutions based on related exercises and the second search dataset.

3. A label matching system for smart teaching, characterized by: The system comprises: An original label storage module is used to store original marking labels for exercises, wherein the original marking labels are labels for the exercises based on the knowledge points and the exercises matched with the knowledge points; A custom tag marking module is used to obtain custom tags during the exercise process and perform custom tags on the exercises according to the custom tags, wherein the custom tags include custom semantic tags and custom keyword tags; A retrieval module is used to obtain a question retrieval formula during the question retrieval process, wherein the question retrieval formula includes at least a standard retrieval term corresponding to the original label and a custom retrieval term corresponding to the custom label; A solution recommendation module is used to generate a solution recommendation for the exercise based on the exercise search formula, the custom tag and the original tag; Methods for obtaining custom labels include: Obtaining custom options, wherein the custom options include semantic options and keyword options; Generating the custom label according to the custom options and custom entry information; The method for generating a custom label based on the custom options and custom input information includes: When the custom option is a semantic option, identifying multiple semantic keywords in the custom input information, and generating a custom semantic tag based on the current exercise content and the multiple semantic keywords; When the custom option is a keyword option, the custom keyword in the custom entry information is used as the custom keyword label; The method of identifying multiple semantic keywords in the custom input information and generating a custom semantic tag based on the current exercise content and the multiple semantic keywords includes: Identifying standard words in the custom search terms based on the current exercise content; Obtaining the plurality of semantic keywords from the selected standard words and the entered supplementary words; Arranging the multiple semantic keywords into logical relationships to obtain the custom semantic tag; The selected standard words are obtained by students choosing from the standard words; Arranging the logical relationships of multiple semantic keywords to obtain the custom semantic tag includes: The students logically sort the multiple semantic keywords and add logical conjunctions between the semantic keywords to obtain the custom semantic labels; The system further comprises: Storing the custom semantic tags and custom keyword tags in a semantic tag set and a keyword tag set respectively; When obtaining a question search formula, searching for matching custom tags in a semantic tag set or a keyword tag set based on the custom search term and the selected custom classification, wherein the custom classification includes a semantic class corresponding to the semantic tag set and a keyword class corresponding to the keyword tag set; The method for the custom tag marking module to custom tag the exercises according to the custom tag includes performing a logical relationship analysis on each exercise using a logical analysis model, analyzing the logical position of the custom keyword in the current exercise, searching for exercises containing the custom keyword and having the same logical position of the custom keyword in the exercise, and marking the retrieved exercises with the custom keyword tag: When the custom tag is a custom keyword tag, identifying the logical position of the custom keyword in the current exercise, and marking the exercise containing the custom keyword and having the same logical position of the custom keyword in the exercise with the custom keyword as the custom keyword tag; When the custom tag is a custom semantic tag, the exercises that simultaneously have the multiple semantic keywords and each of which has the same logical position as the semantic keyword in the current exercise are marked with the custom semantic tag; The logical analysis model is a semantic segmentation model. Existing exercises are used as training samples. After the training samples are classified according to logical relationship categories, the content of the samples is divided into logical positions. The samples are input into the semantic segmentation model for training to obtain the logical analysis model. The logical analysis model is used to analyze the logical relationships of each exercise. The logical relationship categories include total-to-specific, primary-secondary, parallel, progressive, point-to-surface, cause-and-effect, qualitative and quantitative. After the logical relationship categories of the exercises are determined, the exercises are divided according to their logical positions. When entering custom search terms, the standard terms are displayed to students in a floating window, which contains the logical relationship category and logical position of the current exercise. When students logically sort multiple semantic keywords, the corresponding semantic keywords are placed in the matching logical positions and then the semantic keywords in the corresponding logical positions are logically sorted.

Images