A hierarchical teaching-oriented intelligent homework pushing method and device

Abstract

The application belongs to the field of intelligent education, and particularly relates to an intelligent homework pushing method and device for hierarchical teaching. The method comprises the following steps: S1, obtaining historical answering data of all questions answered by a user; S2, constructing a time sequence attenuation model, performing differentiated weight assignment on the historical answering data, and calculating an initial mastery index of a target knowledge point; S3, constructing a knowledge point association coupling matrix based on structure association, co-occurrence association and semantic association, correcting the initial mastery index of the target knowledge point according to the knowledge point association coupling matrix, and obtaining a corrected mastery index of the target knowledge point; and S4, performing difficulty level division on all questions corresponding to the target knowledge point in a preset question bank according to the corrected mastery index, and filtering out questions for homework pushing according to the difficulty level. The application realizes precise hierarchical teaching and improves the pertinence of homework pushing and learning effect.

Description

Technical Field

[0001] This application belongs to the field of intelligent education, and specifically relates to an intelligent homework assignment method and device for differentiated instruction. Background Technology

[0002] With the development of educational informatization, intelligent homework assignment systems are increasingly being applied to differentiated instruction scenarios. Current technologies typically determine question difficulty based on students' historical accuracy rates or simple statistical indicators, lacking consideration for the temporal characteristics of answer data. For example, early and recent answer records are given the same weight in mastery assessments, making it difficult for the results to reflect the latest changes in students' knowledge levels. Furthermore, knowledge points often have interrelationships, such as prerequisite and subsequent knowledge, cognate and derived concepts. Existing methods often assess each knowledge point in isolation, failing to consider the coupling effects between them, easily leading to biased mastery assessments. In addition, in the question assignment process, existing solutions often rely on single difficulty tags for filtering, lacking a mechanism for fine-grained difficulty level classification, making it difficult to achieve truly differentiated instruction.

[0003] Therefore, there is an urgent need for a technical solution to overcome or mitigate at least one of the aforementioned defects in the existing technology. Summary of the Invention

[0004] The purpose of this application is to provide an intelligent homework assignment method and apparatus for differentiated instruction, in order to solve at least one problem existing in the prior art.

[0005] The technical solution of this application is:

[0006] The first aspect of this application provides an intelligent homework assignment method for differentiated instruction, including:

[0007] S1. Obtain historical answer data for all questions answered by the user;

[0008] S2. Construct a time-series decay model, assign differentiated weights to the historical answer data, and calculate the initial mastery index of the target knowledge point;

[0009] S3. Construct a knowledge point association coupling matrix based on structural association, co-occurrence association, and semantic association. Correct the initial mastery index of the target knowledge point according to the knowledge point association coupling matrix to obtain the corrected mastery index of the target knowledge point.

[0010] S4. Based on the revised mastery index, classify all questions corresponding to the target knowledge points in the preset question bank into difficulty levels, and select questions based on the difficulty levels for homework assignment.

[0011] In at least one embodiment of this application, in S1, the historical answer data includes answer duration, answer correctness result, answer time, and knowledge point tag.

[0012] In at least one embodiment of this application, in S2, a time-series decay model is constructed, and differentiated weights are assigned to the historical answer data to calculate the initial mastery index of the target knowledge point, including:

[0013] S21. Construct a time-series decay model:

[0014] ;

[0015] in, Let λ be the temporal decay weight of the i-th question corresponding to the target knowledge point, and λ be the knowledge forgetting decay coefficient. This represents the time difference between the current moment and the time of answering the i-th question;

[0016] S22. Calculate the mastery score for the i-th question corresponding to the target knowledge point:

[0017] ;

[0018] in, Let be the score for mastering the i-th question corresponding to the target knowledge point. This is the correction factor for the deviation in answering time. Let i be the answering time for the i-th question. The standard answering time for the i-th question is... The coefficient for incorrect answers;

[0019] S23. Calculate the initial mastery index of the target knowledge points:

[0020] ;

[0021] in, is the initial mastery index of the target knowledge point, and n is the total number of questions corresponding to the target knowledge point.

[0022] In at least one embodiment of this application, in step S3, a knowledge point association coupling matrix is ​​constructed based on structural association, co-occurrence association, and semantic association. The initial mastery index of the target knowledge point is then corrected according to the knowledge point association coupling matrix to obtain the corrected mastery index of the target knowledge point, including:

[0023] S31. Based on structural association, co-occurrence association, and semantic association, obtain the coupling weight coefficients of all associated knowledge points to the target knowledge point, and construct the knowledge point association coupling matrix.

[0024] S32. Calculate the initial mastery index of all related knowledge points;

[0025] S33. Based on the knowledge point association and coupling matrix and the initial mastery index of all associated knowledge points, the initial mastery index of the target knowledge point is corrected to obtain the corrected mastery index of the target knowledge point:

[0026] ;

[0027] in, The revised mastery index for the target knowledge points. The ontology weight coefficient for the target knowledge point. This represents the coupling weight coefficient between the j-th associated knowledge point and the target knowledge point in the knowledge point association coupling matrix. Let be the initial mastery index for the j-th related knowledge point, and k be the total number of related knowledge points.

[0028] In at least one embodiment of this application, in step S31, the coupling weight coefficients of all associated knowledge points to the target knowledge point are obtained based on structural association, co-occurrence association, and semantic association, and a knowledge point association coupling matrix is ​​constructed, including:

[0029] Extract the node relationships between related knowledge points and target knowledge points from the knowledge graph, and determine the structural association weights based on the node relationships;

[0030] Analyze the co-occurrence frequency of questions related to the target knowledge point in historical answer data, and determine the co-occurrence association weight based on the question co-occurrence frequency.

[0031] The descriptive texts of related knowledge points and target knowledge points are vectorized, and the semantic association weights are determined based on the cosine similarity of the vectors.

[0032] The structural association weight, co-occurrence association weight, and semantic association weight are weighted, summed, and normalized according to a preset fusion coefficient to obtain the coupling weight coefficient of the associated knowledge point to the target knowledge point.

[0033] Construct a knowledge point association coupling matrix based on the coupling weight coefficients of all related knowledge points to the target knowledge point.

[0034] In at least one embodiment of this application, in step S4, the difficulty level of all questions corresponding to the target knowledge point in the preset question bank is divided according to the modified mastery index, including:

[0035] S41. Compare the revised mastery index with the preset index range to determine the mastery level of the target knowledge point.

[0036] S42. Obtain the inherent difficulty coefficient of all questions corresponding to the target knowledge point in the preset question bank;

[0037] S43. Based on the mastery level, set the corresponding difficulty coefficient to divide the range, and divide all questions corresponding to the target knowledge points in the preset question bank into multiple difficulty levels.

[0038] In at least one embodiment of this application, in S43, the difficulty levels include a basic level, an advanced level, and an extended level.

[0039] A second aspect of this application provides an intelligent homework assignment device for differentiated instruction, comprising:

[0040] The historical answer data acquisition module is used to acquire historical answer data for all questions answered by the user.

[0041] The initial mastery index calculation module is used to construct a time-series decay model, assign differentiated weights to the historical answer data, and calculate the initial mastery index of the target knowledge point.

[0042] The mastery index correction module is used to construct a knowledge point association coupling matrix based on structural association, co-occurrence association, and semantic association, and to correct the initial mastery index of the target knowledge point according to the knowledge point association coupling matrix to obtain the corrected mastery index of the target knowledge point.

[0043] The homework assignment module is used to classify all questions corresponding to the target knowledge points in the preset question bank into difficulty levels according to the revised mastery index, and to select questions for homework assignment based on the difficulty levels.

[0044] The invention has at least the following beneficial technical effects:

[0045] The intelligent homework recommendation method for differentiated instruction proposed in this application assigns differentiated weights to historical answer data by constructing a time-series decay model, making the mastery index assessment closer to the user's current actual level; it corrects the initial mastery index through a knowledge point association coupling matrix, effectively eliminating the bias caused by independent assessment of knowledge points; and it divides the difficulty level based on the corrected mastery index and selects and recommends questions, thereby achieving precise differentiated instruction and improving the relevance of homework recommendations and learning effectiveness. Attached Figure Description

[0046] Figure 1 This is a flowchart of an intelligent homework assignment method for tiered teaching, according to one embodiment of this application.

[0047] Figure 2 This is a schematic diagram of an intelligent homework assignment device for tiered teaching according to one embodiment of this application. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, embodiments of this application. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0049] The following is in conjunction with the appendix Figures 1 to 2 This application will be described in further detail.

[0050] The first aspect of this application provides an intelligent homework assignment method for differentiated instruction, comprising the following steps:

[0051] S1. Obtain historical answer data for all questions answered by the user;

[0052] S2. Construct a time-series decay model, assign differentiated weights to historical answer data, and calculate the initial mastery index of the target knowledge points.

[0053] S3. Construct a knowledge point association coupling matrix based on structural association, co-occurrence association, and semantic association. Correct the initial mastery index of the target knowledge point according to the knowledge point association coupling matrix to obtain the corrected mastery index of the target knowledge point.

[0054] S4. Based on the revised mastery index, classify all questions in the preset question bank corresponding to the target knowledge points into difficulty levels, and select questions based on the difficulty level for homework assignment.

[0055] The intelligent homework push method for differentiated instruction proposed in this application firstly obtains the historical answer data of all questions answered by the user in S1, including the answering time, correctness result, answering time, and knowledge point tags for each question, so as to provide data support for the subsequent calculation of the initial mastery index.

[0056] Secondly, in S2, a time-series decay model is constructed based on the forgetting curve. Differential weights are assigned to historical answer data at different times, with higher weights for historical answer data closer to the current time. The weights of historical answer data decay continuously over time. An initial mastery index is calculated based on the time-series decay weights and the mastery score for each question. Specifically, this includes:

[0057] S21. Construct a time-series decay model:

[0058] ;

[0059] in, Let λ be the temporal decay weight of the i-th question corresponding to the target knowledge point, and λ be the knowledge forgetting decay coefficient. This represents the time difference between the current moment and the time of answering the i-th question;

[0060] S22. Calculate the mastery score for the i-th question corresponding to the target knowledge point:

[0061] ;

[0062] in, Let be the score for mastering the i-th question corresponding to the target knowledge point. This is the correction factor for the deviation in answering time. Let i be the answering time for the i-th question. The standard answering time for the i-th question is... The coefficient for incorrect answers;

[0063] S23. Calculate the initial mastery index of the target knowledge points:

[0064] ;

[0065] in, is the initial mastery index of the target knowledge point, and n is the total number of questions corresponding to the target knowledge point.

[0066] By combining the forgetting curve with time-series weighted calculations, the interference of outdated learning data is reduced, effectively improving the timeliness and accuracy of learning assessment for knowledge points.

[0067] The intelligent homework recommendation method for tiered teaching in this application, in step S3, constructs a knowledge point association coupling matrix based on structural association, co-occurrence association, and semantic association, containing coupling weight coefficients of all associated knowledge points to the target knowledge point. The initial mastery index of the target knowledge point is then corrected based on the initial mastery index of the associated knowledge points. The specific process is as follows:

[0068] S31. Based on structural association, co-occurrence association, and semantic association, obtain the coupling weight coefficients of all associated knowledge points to the target knowledge point, and construct the knowledge point association coupling matrix.

[0069] S32. Calculate the initial mastery index of all related knowledge points;

[0070] S33. Based on the knowledge point association and coupling matrix and the initial mastery index of all associated knowledge points, the initial mastery index of the target knowledge point is corrected to obtain the corrected mastery index of the target knowledge point:

[0071] ;

[0072] in, The revised mastery index for the target knowledge points. The ontology weight coefficient for the target knowledge point. This represents the coupling weight coefficient between the j-th associated knowledge point and the target knowledge point in the knowledge point association coupling matrix. Let be the initial mastery index for the j-th related knowledge point, and k be the total number of related knowledge points.

[0073] In a preferred embodiment of this application, the process of constructing the knowledge point association coupling matrix includes:

[0074] Extract the node relationships between related knowledge points and target knowledge points from the knowledge graph, and determine the structural association weights based on the node relationships;

[0075] Analyze the co-occurrence frequency of questions related to the target knowledge point in historical answer data, and determine the co-occurrence association weight based on the question co-occurrence frequency.

[0076] The descriptive texts of related knowledge points and target knowledge points are vectorized, and the semantic association weights are determined based on the cosine similarity of the vectors.

[0077] The structural association weight, co-occurrence association weight, and semantic association weight are weighted, summed, and normalized according to a preset fusion coefficient to obtain the coupling weight coefficient of the associated knowledge point to the target knowledge point.

[0078] Construct a knowledge point association coupling matrix based on the coupling weight coefficients of all related knowledge points to the target knowledge point.

[0079] In this embodiment, the coupling weight coefficients of the associated knowledge points to the target knowledge points are determined from three dimensions: structural association, co-occurrence association, and semantic association.

[0080] Based on the node corresponding to the target knowledge point in the knowledge graph, we retrieve related knowledge point nodes that are directly or indirectly associated with it, and identify the relationships between them, including prerequisite relationships, successor relationships, parent-child relationships, shared parent nodes, and shared child nodes. A prerequisite relationship indicates that mastering the related knowledge point is a prerequisite for learning the target knowledge point; its structural association is usually stronger than that of parallel knowledge points at the same level. A successor relationship indicates that the related knowledge point is a subsequent extension of the target knowledge point. A parent-child relationship indicates a hierarchical inclusion relationship between the related and target knowledge points; the parent node is a higher-level concept or chapter, and the child node is a lower-level concept or sub-knowledge point, reflecting the abstract and concrete affiliation of concepts. A shared parent node indicates that the related and target knowledge points belong to the same higher-level concept or chapter, and are parallel at the same level. A shared child node indicates that the related and target knowledge points jointly contain the same lower-level concept or sub-knowledge point, and there is an intersection between them on that sub-knowledge point. The above structural relationships all reflect the proximity of knowledge points in the graph structure. By assigning corresponding weights to different structural relationships and weighting and fusing them with the shortest path distance of each structural relationship, the structural relationship weights can be obtained.

[0081] The strength of the co-occurrence relationship between related and target knowledge points is measured by statistically analyzing the frequency of their co-occurrence in the same question within historical answer data. A higher co-occurrence frequency indicates a closer connection between the two knowledge points in actual testing, and a greater co-occurrence relationship weight.

[0082] Semantic association is used to quantify the strength of the association between related knowledge points and target knowledge points at the textual semantic level. Descriptive texts of both related and target knowledge points are obtained, including knowledge point names, definitions, chapter descriptions, and teaching requirements. Then, a text vectorization model, such as the bag-of-words model, TF-IDF, or a pre-trained language model, is used to map the descriptive texts into semantic vectors. The cosine similarity between the two vectors is calculated and normalized to the range of 0 to 1, serving as the semantic association weight. A higher cosine similarity indicates a stronger semantic association between the two knowledge points, while a lower cosine similarity indicates a weaker semantic association.

[0083] Finally, the three types of association weights are weighted, summed, and normalized according to a preset fusion coefficient to obtain the coupling weight coefficient of the associated knowledge point to the target knowledge point. The larger the coupling weight coefficient, the stronger the influence of the associated knowledge point on the target knowledge point.

[0084] By considering the coupling effect of related knowledge points on target knowledge points, the limitations of independent evaluation of knowledge points are avoided. This allows for the accurate identification of users' hidden knowledge weaknesses and false mastery, thus making the evaluation results of knowledge point mastery more consistent with their true learning level.

[0085] The intelligent homework assignment method for differentiated instruction in this application, in step S4, categorizes the difficulty levels of questions based on a revised mastery index. The specific process is as follows:

[0086] S41. Compare the revised mastery index with the preset index range to determine the mastery level of the target knowledge point.

[0087] S42. Obtain the inherent difficulty coefficient of all questions corresponding to the target knowledge point in the preset question bank;

[0088] S43. Based on the mastery level, set the corresponding difficulty coefficient to divide the range, and divide all questions corresponding to the target knowledge points in the preset question bank into multiple difficulty levels.

[0089] The revised mastery index is pre-normalized to a value between 0 and 1, and then compared with a preset index range to determine the mastery level of the target knowledge point. The inherent difficulty coefficient of a question is related to factors such as question type complexity, number of nested knowledge points, number of solution steps, the inherent difficulty of the knowledge point itself, historical answer performance, cognitive level requirements, and the design of distraction traps. Taking all the above factors into account, the inherent difficulty coefficient in the range [0, 1] is obtained through pre-trained model or expert annotation and stored when the question is added to the database.

[0090] By combining the revised mastery index with the inherent difficulty coefficient of the questions, an adaptive hierarchical classification mechanism for user learning and question difficulty is constructed. The threshold of the difficulty coefficient range is adjusted according to the mastery level, and all questions corresponding to the target knowledge points in the preset question bank are divided into basic level, intermediate level and advanced level.

[0091] In one specific embodiment of this application, when the modified mastery index satisfies 0 ≤ M < 0.3, the mastery level is considered weak. Questions with an inherent difficulty coefficient in the range [0, 0.2) are classified as basic questions, those in the range [0.2, 0.6) as advanced questions, and those in the range [0.6, 1.0] as extended questions. When the modified mastery index satisfies 0.3 ≤ M < 0.7, the mastery level is considered moderate. Questions with an inherent difficulty coefficient in the range [0, 0.3) are classified as basic questions, those in the range [0.3, 0.7) as advanced questions, and those in the range [0.7, 1.0] as extended questions. When the revised mastery index satisfies 0.7≤M≤1.0, the mastery level is good. Questions with inherent difficulty coefficients in the range of [0, 0.4) are classified as basic questions, questions with inherent difficulty coefficients in the range of [0.4, 0.8) are classified as advanced questions, and questions with inherent difficulty coefficients in the range of [0.8, 1.0] are classified as extension questions.

[0092] Through the aforementioned adaptive hierarchical classification mechanism based on user learning progress and question difficulty, the difficulty level of questions is dynamically matched with the user's actual mastery of the target knowledge points. This avoids frustration for users with weaker learning progress who are directly exposed to high-difficulty questions, and also prevents inefficient practice of low-difficulty questions by users with good learning progress. As the mastery level increases, the difficulty range of basic, intermediate, and advanced questions shifts upwards, forming a gradual, progressive gradient. Setting difficulty level intervals based on mastery level differences is more targeted and adaptive than fixed difficulty stratification, facilitating differentiated instruction and personalized homework assignments, thereby improving learning effectiveness and motivation.

[0093] This application presents an intelligent homework recommendation method for differentiated instruction. It acquires users' historical answer data and constructs a time-series decay model, assigning differentiated weights to the historical answer data to calculate the initial mastery index of the target knowledge point, making the mastery assessment more aligned with the current learning situation. Furthermore, it constructs a knowledge point association coupling matrix, integrating structural associations, co-occurrence associations, and semantic associations to correct the initial mastery index, resulting in a more accurate revised mastery index. Finally, it constructs an adaptive hierarchical classification mechanism based on user learning situation and question difficulty, dividing questions into basic, advanced, and extended questions, and then filtering and recommending them. This method achieves dynamic and precise matching of question difficulty with the user's actual mastery, accurately adapting to the differentiated instruction needs of users with different learning situations.

[0094] Based on the above-described intelligent homework assignment method for differentiated instruction, a second aspect of this application provides an intelligent homework assignment device for differentiated instruction, comprising:

[0095] The historical answer data acquisition module is used to acquire historical answer data for all questions answered by the user.

[0096] The initial mastery index calculation module is used to construct a time-series decay model, assign differentiated weights to historical answer data, and calculate the initial mastery index of the target knowledge point.

[0097] The mastery index correction module is used to construct a knowledge point association coupling matrix based on structural association, co-occurrence association, and semantic association. The initial mastery index of the target knowledge point is corrected according to the knowledge point association coupling matrix to obtain the corrected mastery index of the target knowledge point.

[0098] The homework assignment module is used to classify all questions corresponding to the target knowledge points in the preset question bank into difficulty levels based on the revised mastery index, and then select questions based on the difficulty level for homework assignment.

[0099] The intelligent homework assignment push device for differentiated instruction in this application is designed in detail for each module with reference to the intelligent homework assignment push method for differentiated instruction, which will not be described in detail here.

[0100] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for intelligent homework assignment delivery for differentiated instruction, characterized in that, include: S1. Obtain historical answer data for all questions answered by the user; S2. Construct a time-series decay model, assign differentiated weights to the historical answer data, and calculate the initial mastery index of the target knowledge point; S3. Construct a knowledge point association coupling matrix based on structural association, co-occurrence association, and semantic association. Correct the initial mastery index of the target knowledge point according to the knowledge point association coupling matrix to obtain the corrected mastery index of the target knowledge point. S4. Based on the revised mastery index, classify all questions corresponding to the target knowledge points in the preset question bank into difficulty levels, and select questions based on the difficulty levels for homework assignment.

2. The intelligent homework assignment method for differentiated instruction according to claim 1, characterized in that, In S1, the historical answer data includes answer duration, answer correctness result, answer time, and knowledge point tags.

3. The intelligent homework assignment method for differentiated instruction according to claim 2, characterized in that, In S2, a time-series decay model is constructed to assign differentiated weights to the historical answer data and calculate the initial mastery index of the target knowledge point, including: S21. Construct a time-series decay model: ； in, Let λ be the temporal decay weight of the i-th question corresponding to the target knowledge point, and λ be the knowledge forgetting decay coefficient. This represents the time difference between the current moment and the time of answering the i-th question; S22. Calculate the mastery score for the i-th question corresponding to the target knowledge point: ； in, Let be the score for mastering the i-th question corresponding to the target knowledge point. This is the correction factor for the deviation in answering time. Let i be the answering time for the i-th question. The standard answering time for the i-th question is... The coefficient for incorrect answers; S23. Calculate the initial mastery index of the target knowledge points: ； in, is the initial mastery index of the target knowledge point, and n is the total number of questions corresponding to the target knowledge point.

4. The intelligent homework assignment method for differentiated instruction according to claim 3, characterized in that, In S3, a knowledge point association coupling matrix is ​​constructed based on structural association, co-occurrence association, and semantic association. The initial mastery index of the target knowledge point is then corrected based on this knowledge point association coupling matrix to obtain the corrected mastery index of the target knowledge point, including: S31. Based on structural association, co-occurrence association, and semantic association, obtain the coupling weight coefficients of all associated knowledge points to the target knowledge point, and construct the knowledge point association coupling matrix. S32. Calculate the initial mastery index of all related knowledge points; S33. Based on the knowledge point association and coupling matrix and the initial mastery index of all associated knowledge points, the initial mastery index of the target knowledge point is corrected to obtain the corrected mastery index of the target knowledge point: ； in, The revised mastery index for the target knowledge points. The ontology weight coefficient for the target knowledge point. This represents the coupling weight coefficient between the j-th associated knowledge point and the target knowledge point in the knowledge point association coupling matrix. Let be the initial mastery index for the j-th related knowledge point, and k be the total number of related knowledge points.

5. The intelligent homework assignment method for differentiated instruction according to claim 4, characterized in that, In S31, the coupling weight coefficients of all associated knowledge points to the target knowledge point are obtained based on structural association, co-occurrence association, and semantic association, and a knowledge point association coupling matrix is ​​constructed, including: Extract the node relationships between related knowledge points and target knowledge points from the knowledge graph, and determine the structural association weights based on the node relationships; Analyze the co-occurrence frequency of questions related to the target knowledge point in historical answer data, and determine the co-occurrence association weight based on the question co-occurrence frequency. The descriptive texts of related knowledge points and target knowledge points are vectorized, and the semantic association weights are determined based on the cosine similarity of the vectors. The structural association weight, co-occurrence association weight, and semantic association weight are weighted, summed, and normalized according to a preset fusion coefficient to obtain the coupling weight coefficient of the associated knowledge point to the target knowledge point. Construct a knowledge point association coupling matrix based on the coupling weight coefficients of all related knowledge points to the target knowledge point.

6. The intelligent homework assignment method for differentiated instruction according to claim 5, characterized in that, In S4, based on the revised mastery index, all questions corresponding to the target knowledge points in the preset question bank are divided into difficulty levels, including: S41. Compare the revised mastery index with the preset index range to determine the mastery level of the target knowledge point. S42. Obtain the inherent difficulty coefficient of all questions corresponding to the target knowledge point in the preset question bank; S43. Based on the mastery level, set the corresponding difficulty coefficient to divide the range, and divide all questions corresponding to the target knowledge points in the preset question bank into multiple difficulty levels.

7. The intelligent homework assignment method for differentiated instruction according to claim 6, characterized in that, In S43, the difficulty levels include basic level, advanced level, and extended level.

8. An intelligent homework assignment delivery device for differentiated instruction, characterized in that, include: The historical answer data acquisition module is used to acquire historical answer data for all questions answered by the user. The initial mastery index calculation module is used to construct a time-series decay model, assign differentiated weights to the historical answer data, and calculate the initial mastery index of the target knowledge point. The mastery index correction module is used to construct a knowledge point association coupling matrix based on structural association, co-occurrence association, and semantic association, and to correct the initial mastery index of the target knowledge point according to the knowledge point association coupling matrix to obtain the corrected mastery index of the target knowledge point. The homework assignment module is used to classify all questions corresponding to the target knowledge points in the preset question bank into difficulty levels according to the revised mastery index, and to select questions for homework assignment based on the difficulty levels.

Images