Knowledge graph and teaching content one-key automatic association method and system

Abstract

The invention discloses an intelligent teaching platform knowledge graph and teaching content one-key automatic association method and system, relates to the technical field of intelligent education, and aims at solving the problems that an existing platform is low in association efficiency, single in dimension, poor in visualization and the like. The method comprises the following steps: acquiring and preprocessing a knowledge graph, chapters and job data through one-key triggering; calling an intelligent engine containing four types of matchers (key words, semantics, structures and behaviors), and calculating comprehensive confidence according to weights; establishing association according to a threshold value, detecting conflicts and generating a quality report; updating knowledge graph storage and index; and visually displaying the associated content and the learning condition. The system comprises a one-key trigger module, an intelligent analysis module and the like. According to the method, the association efficiency is improved by hundreds of times, the association quality is guaranteed, dynamic optimization and personalized recommendation are supported, the maintenance cost is reduced, and precise teaching and autonomous learning are assisted.

Description

Technical Field

[0001] This invention relates to the field of smart education technology, specifically to a method and system for automatically associating knowledge graphs with teaching content with a single click. Background Technology

[0002] As educational informatization evolves towards a more intelligent stage, knowledge graphs, with their ability to structurally represent course knowledge points and the logical relationships between them, have become one of the core technologies for building the knowledge system of intelligent teaching platforms. In intelligent teaching scenarios, knowledge graphs can clearly organize the knowledge framework of a course, such as clarifying the prerequisite dependencies between "derivative" and "differential" in "advanced mathematics," or the logical connections between "loop structure" and "conditional judgment" in "programming languages." This provides students with a systematic knowledge cognition path and also provides basic data support for teachers to conduct differentiated instruction and targeted tutoring.

[0003] However, current mainstream smart teaching platforms still face significant technical bottlenecks in the linking of knowledge graphs and teaching content, making it difficult to meet the needs of large-scale and refined teaching. Specific problems are as follows:

[0004] 1. High cost and extremely low efficiency of manual association: In existing platforms, the association of knowledge points in knowledge graphs with teaching content such as chapter courseware, homework, and test questions mainly relies on manual operation by teachers. Taking a course containing 300 core knowledge points as an example, teachers need to match the corresponding teaching resources for each knowledge point one by one. They not only need to confirm the relevance between knowledge points and content, but also need to organize the notes on the association relationship. A single association task often takes 20-30 hours. For scenarios where course resources are frequently updated (such as adjusting 10%-20% of teaching content each semester), teachers need to repeat the association operation, which is extremely burdensome and difficult to adapt to the needs of rapid iteration of teaching resources.

[0005] 2. Limited Dimensions of Association, Insufficient Accuracy and Comprehensiveness: Existing automatic association technologies are mostly limited to keyword matching, that is, comparing high-frequency words in teaching content with extracted knowledge points, lacking in-depth analysis of the semantic connotation and logical structure of the content. For example, in the "Data Structures" course, the knowledge point of "stack" can be associated with courseware content containing the "last-in, first-out" expression through keyword matching, but it cannot be effectively matched with content explaining "function call process" (which implicitly applies the stack), resulting in the omission of a large number of implicit semantic associations. At the same time, relying solely on keywords is prone to misassociations of "same words with different meanings," such as "interfaces in Java" and "network interfaces" being incorrectly associated because both contain the keyword "interface," making it difficult to guarantee the accuracy of association.

[0006] 3. Fragmented System Modules and Lack of a Unified Linkage Mechanism: In most smart teaching platforms, the knowledge graph system, content management system (storing chapter content), and homework system (storing homework questions) operate independently. Each system uses different data formats and storage methods, and no intelligent cross-system linking channel has been established. For example, updating the knowledge point "solving quadratic equations" in the knowledge graph system cannot simultaneously trigger updates to the corresponding courseware in the content management system and the corresponding practice questions in the homework system. This requires manual operation in multiple systems one by one, making it difficult to maintain the consistency and timeliness of the linkage relationships.

[0007] 4. Static and fixed relationships, unable to be dynamically optimized: Once established, existing relationships remain fixed and cannot be adjusted based on teaching feedback and student performance. For example, if most students answer a homework question about "Newton's laws of motion" incorrectly, and analysis reveals that the question actually tests "conservation of momentum" (a deviation from the original related knowledge point), the system cannot automatically correct the relationship between the homework question and the knowledge point. Furthermore, it cannot optimize the relationship logic based on students' learning paths (e.g., students are more likely to grasp "mathematical induction" through learning about "sequences"), leading to a disconnect between the relationships and actual teaching needs.

[0008] In recent years, the rapid development of Natural Language Processing (NLP), graph databases, and multimodal learning technologies has provided key technical support for solving the aforementioned problems. Firstly, pre-trained language models, such as BERT, can deeply understand the semantic information of teaching texts and accurately identify implicit semantic relationships between knowledge points and content, breaking through the limitations of traditional keyword matching. Secondly, graph databases such as Neo4j have the ability to efficiently store and query complex relationships, supporting structured storage of multiple types of relationships (such as "coverage," "examination," and "dependency") between knowledge graph nodes (knowledge points) and teaching content nodes (chapter, assignment), providing a data foundation for cross-system association. Thirdly, multimodal learning technologies can process various forms of teaching content, such as text, images (e.g., formulas and charts), and videos (e.g., experimental demonstrations), achieving effective association between non-textual content and knowledge points. Based on these technological advancements, there is an urgent need for a method and system that can achieve one-click automatic association between knowledge graphs and teaching content, multi-dimensional intelligent matching, dynamic optimization, and visual presentation, in order to break through existing technological bottlenecks and promote the development of smart teaching platforms towards higher efficiency and higher quality intelligence. Summary of the Invention

[0009] The technical problem this invention aims to solve is the inefficiency, single dimension, poor visualization, lack of intelligence, and insufficient real-time performance in the existing intelligent teaching platforms that link knowledge graphs with teaching content. Specifically, the linking relies on manual operation by teachers, which is labor-intensive; it is based on simple tag matching, resulting in low accuracy; it cannot intuitively display linked content and learning progress; it cannot automatically generate personalized learning recommendation paths, and the mastery of knowledge points is not intuitive; and the linking relationships cannot be automatically updated after the teaching content is updated.

[0010] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for one-click automatic association between knowledge graphs and teaching content in a smart teaching platform, comprising the following steps:

[0011] (1) One-click triggering and data collection: Users can start the association process by triggering the one-click association button. The system automatically collects course knowledge graph data, chapter content data and homework database data, and preprocesses and standardizes the collected data.

[0012] (2) Multi-strategy intelligent matching analysis: The intelligent matching engine is called, which includes a keyword matcher, a semantic matcher, a structure matcher and a behavior matcher. Each matcher performs matching analysis on the preprocessed "knowledge point-content" data according to a preset weight, outputs a preliminary confidence score, and then the preliminary confidence scores are weighted and fused to obtain a comprehensive association confidence score.

[0013] (3) Association establishment and verification: The comprehensive association confidence is compared with the preset threshold, and the association that meets the threshold requirements is automatically established. At the same time, relationship conflict detection and automatic resolution are performed, and an association quality report is generated.

[0014] (4) Knowledge graph update: Store the established relationships in the graph database, update the node attributes of the knowledge graph, and establish a relationship index;

[0015] (5) Visualization and Interaction: Display related content on the knowledge graph interface, provide the function of viewing related details, and support manual adjustment of the relationship.

[0016] As a further aspect of the present invention: the keyword matcher in step (2) performs matching analysis based on TF-IDF and text similarity, with a weight of 0.25; the semantic matcher achieves deep semantic understanding based on a pre-trained language model of the Transformer architecture, with a weight of 0.35; the structural matcher performs analysis based on the hierarchical relationship of the course knowledge structure, with a weight of 0.20; and the behavior matcher performs matching based on historical data and user behavior patterns, with a weight of 0.20.

[0017] As a further aspect of the present invention: in step (2), the intelligent matching engine performs matching analysis on the "knowledge point-content" data by calling each matcher in parallel or serially, and the weighted fusion obtains the comprehensive association confidence by calculating the weighted average of each preliminary confidence score.

[0018] As a further aspect of the present invention: if the overall association confidence level in step (3) does not reach the preset threshold, the system executes an alternative strategy. If the alternative strategy is successful, an association relationship is established; if the alternative strategy fails, the association is skipped.

[0019] As a further aspect of the present invention, it also includes an association quality assessment and dynamic optimization mechanism, which periodically assesses the quality of established associations and dynamically adjusts the associations and the weights of each matcher based on teaching feedback and student learning performance.

[0020] As a further aspect of the present invention: in step (5), the visualization of the knowledge graph reflects the number of related contents through the size of the nodes, the color of the nodes indicates the learning completion status, the thickness of the connecting lines indicates the strength of the association, and clicking on a node allows you to view the details of the related contents.

[0021] A smart teaching platform knowledge graph and teaching content automatic association system with one click includes a one-click trigger module, an intelligent analysis module, an association processing module, a graph update module, and a visualization display module;

[0022] The one-click trigger module is used to receive the user's one-click association operation command and trigger the data collection process;

[0023] The intelligent analysis module is used to perform multi-strategy intelligent matching analysis on the collected and preprocessed "knowledge point-content" data and calculate the comprehensive association confidence.

[0024] The association processing module is used to establish association relationships based on comprehensive association confidence, perform conflict detection and resolution, and generate an association quality report;

[0025] The graph update module is used to store associations, update node attributes, and establish association indexes in the graph database;

[0026] The visualization module is used to display related content and learning information on the knowledge graph interface, and provides functions for viewing details and manual adjustment.

[0027] As a further aspect of the present invention: the intelligent analysis module includes a keyword matching unit, a semantic matching unit, a structure matching unit, and a behavior matching unit, which respectively implement keyword matching, semantic matching, structure matching, and behavior matching functions, and is configured with a weighted calculation unit for fusing the preliminary confidence scores output by each matching unit.

[0028] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1-6.

[0029] An electronic device includes a processor and a memory, the memory storing a computer program, wherein the processor executes the computer program to implement the steps of the method according to any one of claims 1-6.

[0030] Compared with the prior art, the present invention has the following advantages:

[0031] 1. Significantly improves the efficiency of linking knowledge points with teaching content (chapter content, homework questions), and greatly reduces the workload of teachers: This invention initiates an automatic linking process through a "one-click trigger" mechanism, replacing the traditional manual operation of linking knowledge points with teaching content (chapter content, homework questions) one by one, achieving an order-of-magnitude improvement in linking efficiency. For a large course containing 300 knowledge points, traditional manual linking requires teachers to invest 20-30 hours, while this invention can complete the entire linking task in just a few minutes; the linking time for small courses is reduced from 5 hours to seconds, and for medium-sized courses from 10-15 hours to minutes, with efficiency improved by up to a hundredfold. This greatly reduces the time teachers spend on mechanical linking work, allowing them to focus more on teaching content design and student tutoring.

[0032] 2. Improve the quality of associations and ensure their accuracy and comprehensiveness: Leveraging a multi-strategy weighted matching algorithm, this approach integrates keyword matching (based on TF-IDF and text similarity), semantic matching (based on deep semantic understanding using a BERT pre-trained model), structural matching (based on hierarchical relationships within the course knowledge hierarchy), and behavioral matching (based on historical data and user behavior patterns), overcoming the limitations of traditional single-keyword matching. On one hand, semantic and structural analysis reduces false associations caused by homonyms, significantly improving association accuracy. On the other hand, behavioral matching and deep semantic mining uncover numerous implicit associations (such as teaching content that implicitly applies knowledge points), drastically reducing the association omission rate. Simultaneously, association quality is quantified using confidence levels, significantly increasing the proportion of high-confidence, high-quality associations and drastically decreasing the proportion of low-confidence, uncertain associations, thus reducing the workload of manual review by teachers.

[0033] 3. Enhance the visual and interactive experience to support teaching and learning decisions: The knowledge graph interface provides an intuitive presentation of relationships and learning data. Node size reflects the amount of related content (e.g., larger nodes indicate more courseware and assignments associated with that knowledge point), node color indicates learning completion status (e.g., green indicates most students have mastered the knowledge, red indicates a low mastery rate), and line thickness indicates the strength of the association (e.g., thicker lines indicate higher confidence in the association between the knowledge point and the content). Clicking on a node allows you to view details of the associated content. Teachers can quickly grasp the coverage of teaching resources for each knowledge point and students' learning progress, allowing for targeted adjustments to teaching focus. Students can clearly understand the structure of the knowledge system and their own learning weaknesses, providing direction for self-directed learning.

[0034] 4. Achieve intelligent recommendation and dynamic optimization to support precise teaching: Based on the relationship between knowledge graph and teaching content, and combined with students' learning performance (such as homework completion and mastery of knowledge points), the system can automatically generate personalized learning recommendation paths (such as recommending corresponding courseware and targeted homework for students who have not mastered "solving quadratic equations in one variable"), solving the "one-size-fits-all" problem in traditional teaching. At the same time, through the association quality assessment and dynamic optimization mechanism, the system regularly updates the association relationship and the weight of each matcher by combining teaching feedback (such as teachers' opinions on adjusting the association relationship) and student behavior data (such as the actual knowledge points tested in a homework question and the deviation of students' answers), so that the association logic continuously meets the actual needs of teaching and ensures the timeliness and applicability of the association relationship.

[0035] 5. Reduced system maintenance costs and improved platform compatibility and scalability: This invention supports automatic synchronization and updating of relationships after teaching content updates, eliminating the need for manual full-scale re-association operations and significantly reducing system maintenance costs. Furthermore, built on a mature technology stack (such as the BERT model and Neo4j graph database), it is compatible with the technical architecture of most existing smart teaching platforms, requiring no large-scale modifications to existing platforms during deployment and simplifying maintenance. In addition, the system supports the association of multimodal teaching content (text, images, videos, etc.) with knowledge graphs, and can expand as the forms of teaching resources become richer, possessing excellent scalability and adapting to the future development needs of smart education. Attached Figure Description

[0036] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0037] Figure 1 This is an overall architecture diagram of the intelligent teaching platform system for automatically associating knowledge graphs with teaching content with one click, as per the present invention.

[0038] Figure 2 This is a flowchart of the multi-strategy intelligent matching algorithm of the present invention. Detailed Implementation

[0039] The following is in conjunction with the appendix Figure 1 (System Architecture Diagram) and Figure 2 (Flowchart of intelligent matching algorithm) Taking the "High School Mathematics Compulsory 1" course as an application scenario, this invention provides a detailed description of the specific implementation method and system for automatically associating the knowledge graph and teaching content of the intelligent teaching platform with one click.

[0040] I. System Deployment and Environment Configuration

[0041] This embodiment implements system deployment based on the disclosed technology stack, ensuring the coordinated operation of each module. The specific configuration is as follows:

[0042] 1. Hardware Environment: The server is configured to support parallel data processing, meeting the computing power requirements of multi-strategy matching algorithms; the client (teacher / student terminal) supports mainstream operating systems (Windows 10 and above, macOS 12 and above) and browsers (Chrome 90+, Edge 90+), ensuring smooth interaction of the knowledge graph visualization interface.

[0043] 2. Software environment:

[0044] Database: A graph database (such as Neo4j) is used to store knowledge graph nodes (knowledge points) and relationships (knowledge points - teaching content). A relational database stores user information, basic course data and related quality assessment results. The storage system stores unstructured content such as chapter courseware (PDF) and homework questions. The caching system temporarily stores frequently accessed related data to improve response speed.

[0045] The core algorithms rely on: semantic matching based on pre-trained language models such as BERT, keyword matching based on TF-IDF and text similarity, structural matching based on the analysis of course knowledge hierarchy relationships, and behavioral matching based on historical data and user behavior patterns.

[0046] System framework: The backend implements interface development to support data interaction, the frontend uses visualization technology (such as ECharts) to display the knowledge graph, and the frontend and backend interact with data through a security protocol to ensure data transmission security.

[0047] II. Method Implementation Steps

[0048] (I) Step 1: One-click triggering and data acquisition

[0049] 1. Triggering operation: Teachers log in to the "High School Mathematics Compulsory 1" course management page of the smart teaching platform, click the "Knowledge Graph - One-click association of teaching content" button, and the system pops up a window to prompt the association risk (such as "will cover the existing unlocked association relationship"). After the teacher confirms, the association process is triggered.

[0050] 2. Data Acquisition: The system automatically collects three types of core data through the data service layer, and performs preprocessing and standardization after collection.

[0051] Knowledge graph data: Read the course knowledge point structure data from the graph database, including knowledge point ID, name, parent node ID, and knowledge point description (such as "function monotonicity: understand the definition and master the judgment method"), and ensure that the data format is consistent.

[0052] Chapter content data: Read chapter courseware data from the storage system, extract text content, perform text conversion on images (such as formulas and charts) (such as OCR recognition), remove special characters and extra spaces, and standardize the format of mathematical symbols (such as converting "x^2" to "x"). 2 ”).

[0053] Homework database data: Read homework questions from a relational database, including homework ID, question text, question type, and assessment objective. Filter out duplicate and invalid questions (such as those with missing question stems), and add a course ID to all data to ensure consistent identification across data sources.

[0054] (II) Step 2: Multi-strategy intelligent matching analysis

[0055] This step invokes the disclosed IntelligentMatchingEngine to execute... Figure 2 The matching process shown analyzes the two types of association pairs: "knowledge point - chapter content" and "knowledge point - homework question".

[0056] 1. Initialize the intelligent matching engine: Load the four defined matchers and preset weights. The program code is as follows:

[0057] class IntelligentMatchingEngine:

[0058] def__init__(self):

[0059] self.matchers = {

[0060] 'keyword':KeywordMatcher(weight=0.25), #Keyword matching

[0061] 'semantic':SemanticMatcher(weight=0.35), #Semantic matching

[0062] 'structural':StructuralMatcher(weight=0.20), #Structural matching

[0063] 'behavioral':BehavioralMatcher(weight=0.20)#Behavioral matching

[0064] }

[0065] 2. Parallel Multi-Strategy Analysis: Based on the knowledge point "monotonicity of functions" and the chapter content "1.3 Basic Properties of Functions", and the homework question "Determine whether the function f(x) = x 2 Taking the monotonicity on the interval [0, +∞) as an example, the system calls four matchers in parallel:

[0066] Keyword matching: Based on TF-IDF and text similarity, calculate the keyword intersection rate between knowledge points and content, and output the preliminary confidence level (e.g., 60% intersection rate between knowledge points and chapter content, confidence level 0.6; 100% intersection rate with homework questions, confidence level 0.95).

[0067] Semantic matching: Based on the deep semantic understanding of the BERT pre-trained model, calculate the semantic similarity of the text and output the preliminary confidence score (e.g., similarity between knowledge point and chapter content is 0.88, confidence score is 0.88; similarity between knowledge point and homework question is 0.92, confidence score is 0.92).

[0068] Structure matching: Based on the hierarchical relationship of the course knowledge structure, determine the logical connection between knowledge points and content, and output the preliminary confidence level (e.g., if a chapter is a direct parent content of a knowledge point, the confidence level is 0.9; if the homework question examines a sub-knowledge point, the confidence level is 0.85).

[0069] Behavioral matching: Based on historical association data and user learning behavior patterns, assess the rationality of the association and output preliminary confidence levels (e.g., 85% of teachers have associated the chapter with the knowledge point, the student accuracy rate is 82%, the chapter association confidence level is 0.85; the homework question association confidence level is 0.9).

[0070] 3. Weighted fusion calculation of overall confidence level: The initial confidence level is calculated by weighting the pre-defined weights (keyword 0.25, semantic 0.35, structure 0.20, behavior 0.20).

[0071] The overall confidence level of knowledge points and chapter content = (0.6 × 0.25) + (0.88 × 0.35) + (0.9 × 0.20) + (0.85 × 0.20) = 0.808;

[0072] The overall confidence level of the knowledge points and homework questions = (0.95×0.25)+(0.92×0.35)+(0.85×0.20)+(0.9×0.20)=0.9095.

[0073] 4. Threshold Judgment and Alternative Strategies: A pre-set confidence threshold (e.g., 0.8) is used. If the overall confidence level is greater than or equal to the threshold, the association is established directly. If the threshold is not reached (e.g., the confidence level of a certain association is 0.75), an alternative strategy (e.g., supplementing multimodal feature matching) is executed. If the alternative strategy succeeds, the association is established; otherwise, it is skipped.

[0074] (III) Step 3: Establishing and Verifying Relationships

[0075] 1. Association Establishment: For association pairs with a comprehensive confidence level ≥ the threshold, the system automatically creates an association relationship, generating an association record in the relational database (including association ID, knowledge point ID, associated object type / ID, comprehensive confidence level, and association time). Simultaneously, it creates a structured association relationship in the graph database using disclosed Cypher statements.

[0076] CREATE(chapter:Chapter)-[r:COVERS{confidence:0.808}]->(kp:KnowledgePoint)

[0077] CREATE(assignment:Assignment)-[r:TESTS{confidence:0.9095}]->(kp:KnowledgePoint)

[0078] 2. Conflict Detection and Resolution: The system automatically detects conflicts such as "the same homework question is associated with multiple knowledge points" and "the same knowledge point is associated with duplicate content". If a conflict is detected (e.g., the homework question is associated with a low-confidence knowledge point), the low-confidence association is automatically removed and the high-confidence association is retained; if duplicate associations exist, they are skipped directly to avoid data redundancy.

[0079] 3. Generation of Association Quality Report: After the association process is completed, the system generates an association quality report, which includes efficiency data (e.g., processing 28 knowledge points, 5 chapters, and 120 homework questions, with a total time of 4 minutes and 20 seconds), quality data (e.g., 152 associations were successfully established, with 68% of them being high-confidence associations), and a notification for pending review (no review is required if there are no low-confidence associations, and suggestions are provided if low-confidence associations exist).

[0080] (iv) Step 4: Knowledge Graph Update

[0081] 1. Relationship storage and index creation: While storing relationships in the graph database, add attributes such as "number of associated chapters" and "number of associated assignments" to knowledge point nodes (e.g., "number of associated chapters = 1" and "number of associated assignments = 8" for the "function monotonicity" node), and create a composite index of "knowledge point ID - associated object ID" to improve the speed of relationship query.

[0082] 2. Node attribute update: Based on student learning data (such as homework completion rate and test scores), update the "learning completion rate" attribute of knowledge point nodes (such as "function monotonicity" completion rate of 75%) to provide data support for subsequent visualization and personalized recommendations.

[0083] (V) Step 5: Visualization and Interaction

[0084] 1. Visual presentation: Teachers can view the association results on the knowledge graph interface. The interface provides an intuitive display based on the description: the size of the node reflects the number of associated content (the "function monotonicity" node is larger because it has more associated content), the node color indicates the learning completion status (75% completion rate is light green, and low completion rate is yellow / light red), and the thickness of the line indicates the association strength (high confidence associations have thicker lines).

[0085] 2. Interactive Functions: Supports viewing related details (clicking a node displays a pop-up window showing a list of related content, download links, and homework accuracy), and manually adjusting related relationships (teachers can modify related objects, and the system automatically updates the related relationships and confidence levels), helping teachers flexibly optimize related results.

[0086] III. Dynamic Optimization and Personalized Recommendations

[0087] 1. Dynamic optimization: The system regularly evaluates the quality of the association relationship and adjusts the association relationship based on teaching feedback and student learning performance (e.g., if the confidence level decreases due to a drop in the accuracy of a certain homework question, the association is automatically removed and rematched). At the same time, the weight of the matcher is optimized (e.g., if semantic matching has a high accuracy in formula-type content, the weight can be fine-tuned to 40%).

[0088] 2. Personalized Recommendation: Based on the correlation and student learning data, the recommendation engine automatically generates personalized learning paths. For example, it recommends key paragraphs of related chapters and basic assignments to students with low completion rates for "function monotonicity", and recommends extended assignments and related knowledge points (such as "the relationship between derivatives and function monotonicity") to students with high completion rates.

[0089] IV. Implementation Results Verification

[0090] This embodiment verifies the effectiveness of the invention through application in the "High School Mathematics Compulsory Course 1":

[0091] 1. Efficiency Improvement: Manual association takes 25 hours, while this invention only takes 4 minutes and 20 seconds, improving efficiency by 340 times, which meets the goal of "reducing the association time of large courses to the minute level";

[0092] 2. Quality Assurance: 100% accuracy in correlation, with an 85% increase in implicit correlation coverage (e.g., discovering an implicit correlation between the "function monotonicity" and "quadratic function graph analysis" courseware);

[0093] 3. Teaching aids: Teachers can quickly identify weak areas (such as the low completion rate of "the concept of derivative") through a visual interface, and students can improve their learning efficiency by 20% through personalized recommendations (the accuracy rate of assignments increased from 65% to 78%).

[0094] In summary, this embodiment fully implements the disclosed method for one-click automatic association between knowledge graphs and teaching content, meeting the efficiency, quality, visualization, and intelligent recommendation requirements of smart teaching platforms, and can be extended to other subject teaching scenarios.

[0095] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for one-key automatic association between a knowledge graph and teaching content in a smart teaching platform, characterized in that, Comprise the following steps: (1) One key trigger and data collection: the user starts the associated process by triggering the associated button, the system automatically collects course knowledge graph data, chapter content data and homework library data, and pre-processes and standardizes the collected data; (2) Multi-strategy intelligent matching analysis: call the intelligent matching engine, which includes keyword matcher, semantic matcher, structure matcher and behavior matcher, each matcher respectively matches the pre-processed "knowledge point-content" data according to the preset weight, outputs the preliminary confidence score, and then weights and fuses each preliminary confidence score to get the comprehensive correlation confidence; (3) Association relationship establishment and verification: compare the comprehensive correlation confidence with the preset threshold, automatically establish the correlation relationship that meets the threshold requirement, detect and automatically solve the relationship conflict, and generate the correlation quality report; (4) Knowledge graph update: store the established correlation relationship in the graph database, update the knowledge graph node attribute, and establish the correlation relationship index; (5) Visualization display and interaction: display the associated content in the knowledge graph interface, provide the associated detail viewing function, and support manual adjustment of the associated relationship. 2.The one-key automatic association method of knowledge graph and teaching content in a smart teaching platform according to claim 1, characterized in that, The keyword matcher in step (2) performs matching analysis based on TF-IDF and text similarity, with a weight of 0.25; the semantic matcher realizes deep semantic understanding based on the pre-trained language model of the Transformer architecture, with a weight of 0.35; the structure matcher analyzes based on the hierarchical relationship of course knowledge structure, with a weight of 0.20; the behavior matcher matches according to historical data and user behavior patterns, with a weight of 0.

20. 3.The one-key automatic association method of knowledge graph and teaching content in a smart teaching platform according to claim 1, characterized in that, The intelligent matching engine in step (2) matches the "knowledge point-content" data in parallel or serial manner, and the comprehensive correlation confidence is obtained by calculating the weighted average of each preliminary confidence score. 4.The one-key automatic association method of knowledge graph and teaching content in a smart teaching platform according to claim 1, characterized in that, If the comprehensive correlation confidence does not reach the preset threshold in step (3), the system executes the alternative strategy, and if the alternative strategy is successful, the correlation relationship is established, and if the alternative strategy fails, the correlation is skipped. 5.The one-key automatic association method of knowledge graph and teaching content in a smart teaching platform according to claim 1, characterized in that, It also includes correlation quality evaluation and dynamic optimization mechanism, which regularly evaluates the quality of the established correlation relationship, dynamically adjusts the correlation relationship and the weight of each matcher according to the teaching feedback and student learning performance. 6.The one-key automatic association method of knowledge graph and teaching content in a smart teaching platform according to claim 1, characterized in that, In step (5), the visualization display of the knowledge graph reflects the number of associated content through the node size, the node color represents the learning completion, the line thickness represents the correlation strength, and the associated content details can be viewed by clicking the node.

7. A smart teaching platform knowledge graph and teaching content one-key automatic association system for implementing the method of any one of claims 1-6. Comprise one key trigger module, intelligent analysis module, correlation processing module, graph update module and visualization display module; The one key trigger module is used to receive the one key association operation instruction of the user and trigger the data collection process; The intelligent analysis module is used for multi-strategy intelligent matching analysis of the "knowledge point-content" data collected and pre-processed, and calculation of the comprehensive correlation confidence; The correlation processing module is used to establish the correlation relationship according to the comprehensive correlation confidence, detect and solve the conflict, and generate the correlation quality report; The graph updating module is configured to store the association relationship, update the node attribute, and establish an association relationship index in the graph database. The visual display module is configured to display the associated content and the related learning information on the knowledge graph interface, and provide a detail viewing and manual adjustment function. 8.The one-key automatic association system of knowledge graph and teaching content in a smart teaching platform according to claim 7, characterized in that, The intelligent analysis module includes a keyword matching unit, a semantic matching unit, a structure matching unit, and a behavior matching unit, which respectively correspond to the keyword matching, semantic matching, structure matching, and behavior matching functions, and are configured with a weighted calculation unit for fusing the preliminary confidence scores output by the matching units.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the method in any one of claims 1-6.

10. An electronic device, comprising: The device includes a processor and a memory, and the memory stores a computer program, and the processor executes the computer program to implement the steps of the method in any one of claims 1-6.

Images