Teaching assistance method and system based on artificial intelligence

Abstract

The invention relates to the technical field of teaching assistance. The system comprises a teaching key analysis module, a teaching knowledge graph construction module, a test question screening and distributing module and a teaching quality evaluation module. The teaching key point analysis module is used for identifying and analyzing teaching key points; the teaching knowledge graph construction module analyzes the dependency relationship among the teaching knowledge points according to the multi-source teaching data set to construct a teaching knowledge graph; the test question screening and distributing module is used for screening and distributing test questions in the knowledge base; and the teaching quality evaluation module analyzes the mastering degree of the knowledge points, analyzes and evaluates the teaching quality through the mastering degree and feeds back the teaching quality. According to the invention, the teaching key points and the pre-associated knowledge points are identified, and the test questions are issued according to the classroom interaction condition, so that the question setting quality is ensured, the knowledge points of students are mastered and understood, and the accuracy of teaching quality evaluation is improved.

Description

Technical Field

[0001] This invention relates to the field of teaching aids technology, specifically to a teaching aid method and system based on artificial intelligence. Background Technology

[0002] In modern education systems, examinations, as a standardized assessment tool, occupy a central position. Their quality analysis is not only a crucial link in the teaching process but also an important way to evaluate teachers' teaching quality. By systematically collecting examination data, teachers can accurately diagnose students' weaknesses in knowledge acquisition, providing a scientific basis for subsequent differentiated instruction and personalized tutoring.

[0003] With current technology, selecting test questions solely based on teaching priorities not only fails to ensure the quality of the test questions themselves (such as difficulty, discrimination, and coverage of cognitive levels), but also makes it difficult to reflect individual differences in students' mastery of specific knowledge points due to their uniformity. This masks individual strengths and weaknesses, leading to biases in the assessment of teaching quality based on such test scores. It is also difficult to accurately attribute the causes (such as being unable to distinguish whether the problem lies with teaching, the test questions, or the students themselves), and it cannot provide an effective basis for precise teaching improvement. Summary of the Invention

[0004] The purpose of this invention is to provide a teaching assistance method and system based on artificial intelligence to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: an artificial intelligence-based teaching assistance method applied to an integrated teaching platform. Acquire multidimensional teaching data, and identify and analyze teaching focus based on the multidimensional teaching data, which includes course materials, audio texts of teacher lectures, and classroom interaction data; Multi-source teaching datasets are obtained by acquiring historical teaching data, the dependencies between teaching knowledge points are analyzed based on the multi-source teaching datasets, and a teaching knowledge graph is constructed based on the dependencies. The test questions in the knowledge base are selected and distributed by using teaching focus and teaching knowledge graph respectively; Based on the completion of the distributed test questions, the teaching focus and the mastery of related prerequisite knowledge points are analyzed, and the teaching quality is analyzed, evaluated, and feedback is provided based on the mastery level.

[0006] Furthermore, the specific method for analyzing the dependencies between teaching priorities is as follows: A multi-source teaching dataset is obtained by using historical teaching data. This dataset includes a subset of student answer records and a subset of knowledge points. The student answer record subset includes [a subset of student answer records], and the knowledge point subset includes a knowledge point ID and its corresponding timestamp. Two knowledge points, Za and Zb, are randomly selected from the knowledge point subsets. The dependency direction between the knowledge points is determined based on their corresponding timestamps. If the timestamp of knowledge point Za is earlier than that of knowledge point Zb, then knowledge point Zb may depend on knowledge point Za; conversely, knowledge point Za may depend on knowledge point Zb. When the timestamp of knowledge point Za is earlier than that of knowledge point Zb, the number of questions that correspond to knowledge points Za and Zb is extracted from the subset of student answer records, and the co-occurrence frequency of knowledge points Za and Zb is calculated based on the number of questions. Extract the correct answer rates (Y(Za), Y(Zb), and Y(Za-Zb)) from the subset of student answer records for knowledge points Za and Zb, as well as for questions where both knowledge points appear together. When Y(Za-Zb) is lower than the preset correct answer rate threshold, it indicates that the student's mastery of knowledge point Za or Zb is insufficient. Therefore, it is necessary to analyze and judge the correct answer rates for questions corresponding to the two knowledge points to obtain the dependency between the two knowledge points as [Y(Zb) / Y(Za)], specifically: When Y(Za) exceeds Y(Zb) and Y(Za) and Y(Zb) exceed the preset accuracy threshold, based on the timestamp, it can be determined that knowledge point Zb may depend on Za. Therefore, when Y(Za) exceeds Y(Zb) and Y(Za) and Y(Zb) exceed the preset accuracy threshold, the accuracy of Y(Za-Zb) cannot be lower than the preset accuracy threshold. That is, the degree of dependence between the two knowledge points is determined as [Y(Zb) / Y(Za)]. When Y(Za) is lower than Y(Zb) and Y(Zb) exceeds the preset accuracy threshold, based on the timestamp, knowledge point Zb may depend on Za. Therefore, when Y(Za) is lower than Y(Zb) and Y(Zb) exceeds the preset accuracy threshold, the accuracy of Y(Za-Zb) may be lower than the preset accuracy threshold. That is, the degree of dependence between the two knowledge points is determined as [Y(Zb) / Y(Za)]. The system iterates through the dependencies and degrees between knowledge points, establishes connections between two knowledge points whose dependencies exceed a preset threshold, and assigns values ​​to these connections based on their degree of dependency, thus completing the construction of the teaching knowledge graph.

[0007] Furthermore, the specific method for filtering and distributing test questions in the knowledge base is as follows: The teaching knowledge graph is used to traverse related prerequisite knowledge points according to the dependency direction, focusing on the teaching focus. Test questions are then selected by matching the knowledge points tested in the knowledge base with the teaching focus and related prerequisite knowledge points. This includes: A test question is randomly selected from the knowledge base. The randomly selected test question corresponds to the knowledge point Zn being examined, where Zn represents the i-th knowledge point examined by the randomly selected test question. The applicability of the test question is calculated by comparing the knowledge point Zn with the teaching focus and related prerequisite knowledge points, specifically as follows: It is determined whether a teaching focus exists in knowledge point Zn. If no teaching focus exists in Zn, the applicability of the test question is 0. If a teaching focus exists in Zn, related prerequisite knowledge points are identified in Zn. The degree of dependence between the related prerequisite knowledge points and the teaching focus is used as a weight, and a weighted sum is applied to a preset initial value and divided by the number of related prerequisite knowledge points in Zn. Test questions with applicability exceeding a preset threshold are selected as test questions to be distributed. The interaction level between students and key teaching points is mapped based on multidimensional teaching data. Analysis is conducted through classroom interaction, with specific indicators including the number of questions asked, the quality of answers, and contributions to group discussions, all dynamically weighted and accumulated. The interaction level with key teaching points is ranked as follows: active participation > passive response > silence. The interaction level is tiered, and test questions corresponding to the key teaching points are distributed in reverse order of applicability. Higher interaction levels indicate a greater grasp of the corresponding key teaching points. Therefore, reverse distribution ensures that students, while mastering the key teaching points, do not forget other knowledge points.

[0008] Furthermore, the specific methods for analyzing and evaluating teaching quality are as follows: Based on the knowledge points corresponding to the test questions and the completion status, the mastery of knowledge points is quantitatively analyzed, and a heat map is constructed to visualize the mastery level. A coordinate matrix is ​​established through the knowledge points corresponding to the test questions, with the horizontal axis representing the knowledge points corresponding to the test questions and the vertical axis representing individual students or student groups. According to the preset mastery level threshold, the mastery status of each student for each knowledge point is mapped to a discrete level, including: setting three knowledge point mastery rate level thresholds A, B, and C from high to low, where A, B, and C ∈ (0,1); different mastery levels are assigned differentiated visual codes, where: areas where the mastery rate of teaching focus is lower than B or the mastery rate of related prerequisite knowledge points is lower than C are marked with the first warning color system; areas where the mastery rate of related prerequisite knowledge points belongs to the interval [C, B] or the mastery rate of teaching focus belongs to the interval [B, A) are marked with the second transition color system; and areas where the mastery rate of related prerequisite knowledge points exceeds A or the mastery rate of teaching focus belongs to the interval (B, 1) are marked with the third saturation color system. The teaching quality is determined by analyzing the proportion of different color-coded areas to the total area. When the first warning color-coded area occupies half of the total area, the teaching quality is considered poor. Feedback on the teaching results is then provided so that teachers can conduct targeted instruction based on these results. The teaching result is the students' mastery rate of knowledge points. Based on the students' mastery rate, targeted instruction can be implemented to ensure effective learning and development for students and maximize learning efficiency.

[0009] Furthermore, the specific method for identifying and analyzing teaching focus is as follows: Acquire multidimensional teaching data, and extract key teaching points from different dimensions based on the multidimensional teaching data. The multidimensional teaching data includes course materials, text transcripts of teacher lectures, and classroom interaction data. Extract key teaching points from different dimensions, including: The course materials are assigned initial weights based on the curriculum standards, and the weights are adjusted according to the frequency of cross-chapter references of knowledge points in the materials to obtain the weights of different knowledge points in the course materials, W(Zi), where Zi represents the i-th teaching knowledge point and W(Zi) represents the weight of the i-th knowledge point in the course materials. Based on the BERT model, high-frequency keywords matching knowledge points in course materials are extracted from the audio text of teacher lectures. MFCC features are used to detect speech rate variation segments of high-frequency keywords. Weight enhancement coefficients are applied to the knowledge points associated with sudden drops in speech rate, resulting in the weight U(Zi') of the high-frequency keywords corresponding to the knowledge points in the course materials. Here, Zi' represents the teaching knowledge point in the teacher lecture audio text that matches the teaching knowledge point more than a preset threshold, and U(Zi') represents the weight of the high-frequency keyword corresponding to the teaching knowledge point in the teacher lecture audio text. The high-frequency keyword refers to the keyword whose repetition frequency exceeds the preset threshold. Based on the frequency and accuracy of question responses in classroom interaction data, a weight K(Zi) is generated for different teaching knowledge points through weighted summation. Initial values ​​are assigned to the teaching knowledge points, and the initial values ​​are weighted and summed according to the weights of different teaching knowledge points on different dimensions to obtain the importance of different teaching knowledge points. Specifically, the importance of different teaching priorities is Q(Zi) = Pi × [W(Zi) + U(Zi') + K(Zi)], where Pi is the initial value assigned to different teaching knowledge points, i' ∈ i, and U(Zi') = 0 when i' and i have different teaching knowledge points. Teaching knowledge points whose importance exceeds a preset threshold are marked as teaching priorities.

[0010] An artificial intelligence-based teaching assistance system includes a teaching focus analysis module, a teaching knowledge graph construction module, a test question screening and distribution module, and a teaching quality evaluation module. The teaching focus analysis module is used to identify and analyze teaching focus by acquiring multidimensional teaching data, including course materials, audio texts of teacher lectures, and classroom interaction data. The teaching knowledge graph construction module is used to obtain multi-source teaching datasets through historical teaching data, analyze the dependencies between teaching knowledge points based on the multi-source teaching datasets, and construct a teaching knowledge graph based on the dependencies. The test question filtering and distribution module is used to filter and distribute test questions in the knowledge base based on teaching focus and teaching knowledge graph respectively. The teaching quality assessment module is used to analyze the mastery of teaching focus and related prerequisite knowledge points based on the completion of the distributed test questions, and to analyze, evaluate and provide feedback on the teaching quality based on the mastery level.

[0011] Furthermore, the teaching focus analysis module includes a dependency direction analysis unit and a dependency degree analysis unit; The dependency direction analysis unit is used to obtain a multi-source teaching dataset through historical teaching data. The multi-source teaching dataset includes a subset of student answer records and a subset of knowledge points. Two knowledge points Za and Zb are arbitrarily selected from the subset of knowledge points, and the dependency direction between the knowledge points is determined based on the timestamps corresponding to knowledge points Za and Zb. If the timestamp corresponding to knowledge point Za is earlier than that of knowledge point Zb, then knowledge point Zb may depend on knowledge point Za. Conversely, knowledge point Za depends on knowledge point Zb. The dependency analysis unit is used to extract the number of questions that correspond to knowledge points Za and Zb from the subset of student answer records when the timestamp of knowledge point Za is earlier than that of knowledge point Zb, and to calculate the co-occurrence frequency of knowledge points Za and Zb based on the number of questions. Extract the correct answer rates (Y(Za), Y(Zb), and Y(Za-Zb)) from the subset of student answer records for knowledge points Za and Zb, as well as for questions where both knowledge points appear together. When Y(Za-Zb) is lower than the preset correct answer rate threshold, it indicates that the student's mastery of knowledge point Za or Zb is insufficient. Therefore, it is necessary to analyze and judge the correct answer rates for questions corresponding to the two knowledge points to obtain the dependency between the two knowledge points as [Y(Zb) / Y(Za)], specifically: When Y(Za) exceeds Y(Zb) and Y(Za) and Y(Zb) exceed the preset accuracy threshold, based on the timestamp, it can be determined that knowledge point Zb may depend on Za. Therefore, when Y(Za) exceeds Y(Zb) and Y(Za) and Y(Zb) exceed the preset accuracy threshold, the accuracy of Y(Za-Zb) cannot be lower than the preset accuracy threshold. That is, the degree of dependence between the two knowledge points is determined as [Y(Zb) / Y(Za)]. When Y(Za) is lower than Y(Zb) and Y(Zb) exceeds the preset accuracy threshold, based on the timestamp, knowledge point Zb may depend on Za. Therefore, when Y(Za) is lower than Y(Zb) and Y(Zb) exceeds the preset accuracy threshold, the accuracy of Y(Za-Zb) may be lower than the preset accuracy threshold. That is, the degree of dependence between the two knowledge points is [Y(Zb) / Y(Za)].

[0012] Furthermore, the test question filtering and distribution module includes a question filtering unit and a question distribution unit; The test question filtering unit is used to traverse related prerequisite knowledge points according to the teaching knowledge graph, focusing on the teaching focus and following the dependency direction. It then filters test questions based on matching the knowledge points examined in the knowledge base with the teaching focus and related prerequisite knowledge points. This includes: A test question is randomly selected from the knowledge base. The randomly selected test question corresponds to the knowledge point Zn being examined, where Zn represents the i-th knowledge point examined by the randomly selected test question. The applicability of the test question is calculated by comparing the knowledge point Zn with the teaching focus and related prerequisite knowledge points, specifically as follows: It is determined whether a teaching focus exists in knowledge point Zn. If no teaching focus exists in Zn, the applicability of the test question is 0. If a teaching focus exists in Zn, related prerequisite knowledge points are identified in Zn. The degree of dependence between the related prerequisite knowledge points and the teaching focus is used as a weight, and a weighted sum is applied to a preset initial value and divided by the number of related prerequisite knowledge points in Zn. Test questions with applicability exceeding a preset threshold are selected as test questions to be distributed. The test question distribution unit is used to map the degree of interaction between students and key teaching points through multi-dimensional teaching data. Analysis is conducted through classroom interaction, and specific indicators include dynamically weighted summaries of the number of questions asked, the quality of answers, and contributions to group discussions. The degree of interaction with key teaching points is ranked as follows: active participation > passive response > silence. The interaction level is graded, and test questions corresponding to the key teaching points are distributed in reverse order based on their applicability. A higher level of interaction indicates a greater grasp of the corresponding key teaching point. Therefore, reverse distribution ensures that students, while mastering the key teaching points, do not forget other knowledge points.

[0013] Furthermore, the teaching quality assessment module includes a heatmap construction unit and a teaching quality evaluation unit; The heatmap construction unit is used to quantitatively analyze the mastery of knowledge points based on the knowledge points corresponding to the test questions and the completion status, and to build a heatmap to visualize the mastery level. A coordinate matrix is ​​established using the knowledge points corresponding to the test questions, with the horizontal axis representing the knowledge points and the vertical axis representing individual students or student groups. Based on preset mastery thresholds, each student's mastery status for each knowledge point is mapped to a discrete level, including: setting three mastery level thresholds for knowledge points, A, B, and C, from high to low, where A, B, and C ∈ (0,1); and assigning differentiated visual codes to different mastery levels, where: areas with a mastery rate of teaching focus below B or a mastery rate of related prerequisite knowledge points below C are marked with a first warning color scheme; areas with a mastery rate of related prerequisite knowledge points belonging to the interval [C, B] or a mastery rate of teaching focus belonging to the interval [B, A) are marked with a second transition color scheme; and areas with a mastery rate of related prerequisite knowledge points exceeding A or a mastery rate of teaching focus belonging to the interval (B, 1) are marked with a third saturation color scheme. The teaching quality assessment unit is used to analyze and judge the teaching quality based on the proportion of different color-coded areas to the total area. When the first warning color-coded area occupies half of the total area, the teaching quality is judged to be poor, and the teaching results are fed back so that teachers can conduct targeted teaching based on the results. The teaching results are the students' mastery rate of knowledge points. Based on the students' mastery rate of knowledge points, targeted teaching can be carried out to ensure effective learning and development of students and maximize learning efficiency.

[0014] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: The present invention identifies the teaching focus and related prior knowledge points, completes the screening of test questions, and distributes test questions according to classroom interaction, thereby ensuring the quality of the test questions and the understanding of students' knowledge points, improving the accuracy of such test scores in evaluating teaching quality, and providing an effective basis for precise teaching improvement. Attached Figure Description

[0015] 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: Figure 1 This is a flowchart illustrating the steps of an artificial intelligence-based teaching assistance method according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figure 1 This invention provides a technical solution: an artificial intelligence-based teaching assistance method applied to an integrated teaching platform. Acquire multidimensional teaching data, and identify and analyze teaching focus based on the multidimensional teaching data, which includes course materials, audio texts of teacher lectures, and classroom interaction data; Multi-source teaching datasets are obtained by acquiring historical teaching data, the dependencies between teaching knowledge points are analyzed based on the multi-source teaching datasets, and a teaching knowledge graph is constructed based on the dependencies. The test questions in the knowledge base are selected and distributed by using teaching focus and teaching knowledge graph respectively; Based on the completion of the distributed test questions, the teaching focus and the mastery of related prerequisite knowledge points are analyzed, and the teaching quality is analyzed, evaluated, and feedback is provided based on the mastery level.

[0018] Furthermore, the specific method for analyzing the dependencies between teaching priorities is as follows: A multi-source teaching dataset is obtained by using historical teaching data. The multi-source teaching dataset includes a subset of student answer records and a subset of knowledge points. Two knowledge points, Za and Zb, are randomly selected from the subset of knowledge points. The dependency direction between knowledge points is determined based on the timestamps corresponding to knowledge points Za and Zb. If the timestamp corresponding to knowledge point Za is earlier than that of knowledge point Zb, then knowledge point Zb may depend on knowledge point Za. Conversely, knowledge point Za depends on knowledge point Zb. When the timestamp of knowledge point Za is earlier than that of knowledge point Zb, the number of questions that correspond to knowledge points Za and Zb is extracted from the subset of student answer records, and the co-occurrence frequency of knowledge points Za and Zb is calculated based on the number of questions. Extract the correct answer rates (Y(Za), Y(Zb), and Y(Za-Zb)) from the subset of student answer records for knowledge points Za and Zb, as well as for questions where both knowledge points appear together. When Y(Za-Zb) is lower than the preset correct answer rate threshold, it indicates that the student's mastery of knowledge point Za or Zb is insufficient. Therefore, it is necessary to analyze and judge the correct answer rates for questions corresponding to the two knowledge points to obtain the dependency between the two knowledge points as [Y(Zb) / Y(Za)], specifically: When Y(Za) exceeds Y(Zb) and Y(Za) and Y(Zb) exceed the preset accuracy threshold, based on the timestamp, it can be determined that knowledge point Zb may depend on Za. Therefore, when Y(Za) exceeds Y(Zb) and Y(Za) and Y(Zb) exceed the preset accuracy threshold, the accuracy of Y(Za-Zb) cannot be lower than the preset accuracy threshold. That is, the degree of dependence between the two knowledge points is determined as [Y(Zb) / Y(Za)]. When Y(Za) is lower than Y(Zb) and Y(Zb) exceeds the preset accuracy threshold, based on the timestamp, knowledge point Zb may depend on Za. Therefore, when Y(Za) is lower than Y(Zb) and Y(Zb) exceeds the preset accuracy threshold, the accuracy of Y(Za-Zb) may be lower than the preset accuracy threshold. That is, the degree of dependence between the two knowledge points is determined as [Y(Zb) / Y(Za)]. The system iterates through the dependencies and degrees between knowledge points, establishes connections between two knowledge points whose dependencies exceed a preset threshold, and assigns values ​​to these connections based on their degree of dependency, thus completing the construction of the teaching knowledge graph.

[0019] Furthermore, the specific method for filtering and distributing test questions in the knowledge base is as follows: The teaching knowledge graph is used to traverse related prerequisite knowledge points according to the dependency direction, focusing on the teaching focus. Test questions are then selected by matching the knowledge points tested in the knowledge base with the teaching focus and related prerequisite knowledge points. This includes: A test question is randomly selected from the knowledge base. The randomly selected test question corresponds to the knowledge point Zn being examined, where Zn represents the i-th knowledge point examined by the randomly selected test question. The applicability of the test question is calculated by comparing the knowledge point Zn with the teaching focus and related prerequisite knowledge points, specifically as follows: It is determined whether a teaching focus exists in knowledge point Zn. If no teaching focus exists in Zn, the applicability of the test question is 0. If a teaching focus exists in Zn, related prerequisite knowledge points are identified in Zn. The degree of dependence between the related prerequisite knowledge points and the teaching focus is used as a weight, and a weighted sum is applied to a preset initial value and divided by the number of related prerequisite knowledge points in Zn. Test questions with applicability exceeding a preset threshold are selected as test questions to be distributed. The interaction level between students and key teaching points is mapped based on multidimensional teaching data. Analysis is conducted through classroom interaction, with specific indicators including the number of questions asked, the quality of answers, and contributions to group discussions, all dynamically weighted and accumulated. The interaction level with key teaching points is ranked as follows: active participation > passive response > silence. The interaction level is tiered, and test questions corresponding to the key teaching points are distributed in reverse order of applicability. Higher interaction levels indicate a greater grasp of the corresponding key teaching points. Therefore, reverse distribution ensures that students, while mastering the key teaching points, do not forget other knowledge points.

[0020] Furthermore, the specific methods for analyzing and evaluating teaching quality are as follows: Based on the knowledge points corresponding to the test questions and the completion status, the mastery of knowledge points is quantitatively analyzed, and a heat map is constructed to visualize the mastery level. A coordinate matrix is ​​established through the knowledge points corresponding to the test questions, with the horizontal axis representing the knowledge points corresponding to the test questions and the vertical axis representing individual students or student groups. According to the preset mastery level threshold, the mastery status of each student for each knowledge point is mapped to a discrete level, including: setting three knowledge point mastery rate level thresholds A, B, and C from high to low, where A, B, and C ∈ (0,1); different mastery levels are assigned differentiated visual codes, where: areas where the mastery rate of teaching focus is lower than B or the mastery rate of related prerequisite knowledge points is lower than C are marked with the first warning color system; areas where the mastery rate of related prerequisite knowledge points belongs to the interval [C, B] or the mastery rate of teaching focus belongs to the interval [B, A) are marked with the second transition color system; and areas where the mastery rate of related prerequisite knowledge points exceeds A or the mastery rate of teaching focus belongs to the interval (B, 1) are marked with the third saturation color system. The teaching quality is determined by analyzing the proportion of different color-coded areas to the total area. When the first warning color-coded area occupies half of the total area, the teaching quality is considered poor. Feedback on the teaching results is then provided so that teachers can conduct targeted instruction based on these results. The teaching result is the students' mastery rate of knowledge points. Based on the students' mastery rate, targeted instruction can be implemented to ensure effective learning and development for students and maximize learning efficiency.

[0021] Furthermore, the specific method for identifying and analyzing teaching focus is as follows: Acquire multidimensional teaching data, and extract key teaching points from different dimensions based on the multidimensional teaching data. The multidimensional teaching data includes course materials, text transcripts of teacher lectures, and classroom interaction data. Extract key teaching points from different dimensions, including: The course materials are assigned initial weights based on the curriculum standards, and the weights are adjusted according to the frequency of cross-chapter references of knowledge points in the materials to obtain the weights of different knowledge points in the course materials, W(Zi), where Zi represents the i-th teaching knowledge point and W(Zi) represents the weight of the i-th knowledge point in the course materials. Based on the BERT model, high-frequency keywords matching knowledge points in course materials are extracted from the audio text of teacher lectures. MFCC features are used to detect speech rate variation segments of high-frequency keywords. Weight enhancement coefficients are applied to the knowledge points associated with sudden drops in speech rate, resulting in the weight U(Zi') of the high-frequency keywords corresponding to the knowledge points in the course materials. Here, Zi' represents the teaching knowledge point in the teacher lecture audio text that matches the teaching knowledge point more than a preset threshold, and U(Zi') represents the weight of the high-frequency keyword corresponding to the teaching knowledge point in the teacher lecture audio text. The high-frequency keyword refers to the keyword whose repetition frequency exceeds the preset threshold. Based on the frequency and accuracy of question responses in classroom interaction data, a weight K(Zi) is generated for different teaching knowledge points through weighted summation. Initial values ​​are assigned to the teaching knowledge points, and the initial values ​​are weighted and summed according to the weights of different teaching knowledge points on different dimensions to obtain the importance of different teaching knowledge points. Specifically, the importance of different teaching priorities is Q(Zi) = Pi × [W(Zi) + U(Zi') + K(Zi)], where Pi is the initial value assigned to different teaching knowledge points, i' ∈ i, and U(Zi') = 0 when i' and i have different teaching knowledge points. Teaching knowledge points whose importance exceeds a preset threshold are marked as teaching priorities.

[0022] An artificial intelligence-based teaching assistance system includes a teaching focus analysis module, a teaching knowledge graph construction module, a test question screening and distribution module, and a teaching quality evaluation module. The teaching focus analysis module is used to identify and analyze teaching focus by acquiring multidimensional teaching data, including course materials, audio texts of teacher lectures, and classroom interaction data. The teaching knowledge graph construction module is used to obtain multi-source teaching datasets through historical teaching data, analyze the dependencies between teaching knowledge points based on the multi-source teaching datasets, and construct a teaching knowledge graph based on the dependencies. The test question filtering and distribution module is used to filter and distribute test questions in the knowledge base based on teaching focus and teaching knowledge graph respectively. The teaching quality assessment module is used to analyze the mastery of teaching focus and related prerequisite knowledge points based on the completion of the distributed test questions, and to analyze, evaluate and provide feedback on the teaching quality based on the mastery level.

[0023] Furthermore, the teaching focus analysis module includes a dependency direction analysis unit and a dependency degree analysis unit; The dependency direction analysis unit is used to obtain a multi-source teaching dataset through historical teaching data. The multi-source teaching dataset includes a subset of student answer records and a subset of knowledge points. Two knowledge points Za and Zb are arbitrarily selected from the subset of knowledge points, and the dependency direction between the knowledge points is determined based on the timestamps corresponding to knowledge points Za and Zb. If the timestamp corresponding to knowledge point Za is earlier than that of knowledge point Zb, then knowledge point Zb may depend on knowledge point Za. Conversely, knowledge point Za depends on knowledge point Zb. The dependency analysis unit is used to extract the number of questions that correspond to knowledge points Za and Zb from the subset of student answer records when the timestamp of knowledge point Za is earlier than that of knowledge point Zb, and to calculate the co-occurrence frequency of knowledge points Za and Zb based on the number of questions. Extract the correct answer rates (Y(Za), Y(Zb), and Y(Za-Zb)) from the subset of student answer records for knowledge points Za and Zb, as well as for questions where both knowledge points appear together. When Y(Za-Zb) is lower than the preset correct answer rate threshold, it indicates that the student's mastery of knowledge point Za or Zb is insufficient. Therefore, it is necessary to analyze and judge the correct answer rates for questions corresponding to the two knowledge points to obtain the dependency between the two knowledge points as [Y(Zb) / Y(Za)], specifically: When Y(Za) exceeds Y(Zb) and Y(Za) and Y(Zb) exceed the preset accuracy threshold, based on the timestamp, it can be determined that knowledge point Zb may depend on Za. Therefore, when Y(Za) exceeds Y(Zb) and Y(Za) and Y(Zb) exceed the preset accuracy threshold, the accuracy of Y(Za-Zb) cannot be lower than the preset accuracy threshold. That is, the degree of dependence between the two knowledge points is determined as [Y(Zb) / Y(Za)]. When Y(Za) is lower than Y(Zb) and Y(Zb) exceeds the preset accuracy threshold, based on the timestamp, knowledge point Zb may depend on Za. Therefore, when Y(Za) is lower than Y(Zb) and Y(Zb) exceeds the preset accuracy threshold, the accuracy of Y(Za-Zb) may be lower than the preset accuracy threshold. That is, the degree of dependence between the two knowledge points is [Y(Zb) / Y(Za)].

[0024] Furthermore, the test question filtering and distribution module includes a question filtering unit and a question distribution unit; The test question filtering unit is used to traverse related prerequisite knowledge points according to the teaching knowledge graph, focusing on the teaching focus and following the dependency direction. It then filters test questions based on matching the knowledge points examined in the knowledge base with the teaching focus and related prerequisite knowledge points. This includes: A test question is randomly selected from the knowledge base. The randomly selected test question corresponds to the knowledge point Zn being examined, where Zn represents the i-th knowledge point examined by the randomly selected test question. The applicability of the test question is calculated by comparing the knowledge point Zn with the teaching focus and related prerequisite knowledge points, specifically as follows: It is determined whether a teaching focus exists in knowledge point Zn. If no teaching focus exists in Zn, the applicability of the test question is 0. If a teaching focus exists in Zn, related prerequisite knowledge points are identified in Zn. The degree of dependence between the related prerequisite knowledge points and the teaching focus is used as a weight, and a weighted sum is applied to a preset initial value and divided by the number of related prerequisite knowledge points in Zn. Test questions with applicability exceeding a preset threshold are selected as test questions to be distributed. The test question distribution unit is used to map the degree of interaction between students and key teaching points through multi-dimensional teaching data. Analysis is conducted through classroom interaction, and specific indicators include dynamically weighted summaries of the number of questions asked, the quality of answers, and contributions to group discussions. The degree of interaction with key teaching points is ranked as follows: active participation > passive response > silence. The interaction level is graded, and test questions corresponding to the key teaching points are distributed in reverse order based on their applicability. A higher level of interaction indicates a greater grasp of the corresponding key teaching point. Therefore, reverse distribution ensures that students, while mastering the key teaching points, do not forget other knowledge points.

[0025] Furthermore, the teaching quality assessment module includes a heatmap construction unit and a teaching quality evaluation unit; The heatmap construction unit is used to quantitatively analyze the mastery of knowledge points based on the knowledge points corresponding to the test questions and the completion status, and to build a heatmap to visualize the mastery level. A coordinate matrix is ​​established using the knowledge points corresponding to the test questions, with the horizontal axis representing the knowledge points and the vertical axis representing individual students or student groups. Based on preset mastery thresholds, each student's mastery status for each knowledge point is mapped to a discrete level, including: setting three mastery level thresholds for knowledge points, A, B, and C, from high to low, where A, B, and C ∈ (0,1); and assigning differentiated visual codes to different mastery levels, where: areas with a mastery rate of teaching focus below B or a mastery rate of related prerequisite knowledge points below C are marked with a first warning color scheme; areas with a mastery rate of related prerequisite knowledge points belonging to the interval [C, B] or a mastery rate of teaching focus belonging to the interval [B, A) are marked with a second transition color scheme; and areas with a mastery rate of related prerequisite knowledge points exceeding A or a mastery rate of teaching focus belonging to the interval (B, 1) are marked with a third saturation color scheme. The teaching quality assessment unit is used to analyze and judge the teaching quality based on the proportion of different color-coded areas to the total area. When the first warning color-coded area occupies half of the total area, the teaching quality is judged to be poor, and the teaching results are fed back so that teachers can conduct targeted teaching based on the results. The teaching results are the students' mastery rate of knowledge points. Based on the students' mastery rate of knowledge points, targeted teaching can be carried out to ensure effective learning and development of students and maximize learning efficiency.

[0026] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0027] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An artificial intelligence-based teaching assistance method applied to an integrated teaching platform, characterized in that: Acquire multidimensional teaching data and identify and analyze key teaching points based on the multidimensional teaching data; Multi-source teaching datasets are obtained by acquiring historical teaching data, the dependencies between teaching knowledge points are analyzed based on the multi-source teaching datasets, and a teaching knowledge graph is constructed based on the dependencies. The test questions in the knowledge base are selected and distributed by using teaching focus and teaching knowledge graph respectively; Based on the completion of the distributed test questions, the teaching focus and the mastery of related prerequisite knowledge points are analyzed, and the teaching quality is analyzed, evaluated, and feedback is provided based on the mastery level.

2. The teaching assistance method based on artificial intelligence according to claim 1, characterized in that: The specific method for analyzing the dependencies between teaching priorities is as follows: A multi-source teaching dataset is obtained by acquiring historical teaching data. The multi-source teaching dataset includes a subset of student answer records and a subset of knowledge points. Two knowledge points Za and Zb are randomly selected from the subset of knowledge points. The dependency direction between knowledge points is determined based on the timestamps corresponding to knowledge points Za and Zb. When the timestamp of knowledge point Za is earlier than that of knowledge point Zb, the number of questions that correspond to knowledge points Za and Zb is extracted from the subset of student answer records, and the co-occurrence frequency of knowledge points Za and Zb is calculated based on the number of questions. Extract the correct answer rates of knowledge points Za and Zb, as well as the correct answer rates of the two knowledge points appearing together, from the subset of student answer records. When Y(Za-Zb) is lower than the preset correct answer rate threshold, analyze and judge the correct answer rates of the questions corresponding to the two knowledge points to obtain the degree of dependence of the two knowledge points as [Y(Zb) / Y(Za)]. The system iterates through the dependencies and degrees between knowledge points, establishes connections between two knowledge points whose dependencies exceed a preset threshold, and assigns values ​​to these connections based on their degree of dependency, thus completing the construction of the teaching knowledge graph.

3. The teaching assistance method based on artificial intelligence according to claim 2, characterized in that: The specific method for filtering and distributing test questions in the knowledge base is as follows: The teaching knowledge graph is used to traverse related prerequisite knowledge points according to the dependency direction, focusing on the teaching focus. Test questions are then selected by matching the knowledge points tested in the knowledge base with the teaching focus and related prerequisite knowledge points. This includes: A test question is randomly selected from the knowledge base. The knowledge point Zn corresponding to the randomly selected test question is represented as the i-th knowledge point. The applicability of the test question is calculated by comparing the knowledge point Zn with the teaching focus and related prerequisite knowledge points. Test questions with applicability exceeding a preset threshold are selected as test questions to be distributed. Based on multidimensional teaching data, the degree of interaction between students and teaching focus is mapped, the degree of interaction is graded, and the test questions corresponding to the interactive teaching focus are distributed in reverse order according to their applicability.

4. The teaching assistance method based on artificial intelligence according to claim 3, characterized in that: The specific methods for analyzing and evaluating teaching quality are as follows: Based on the knowledge points corresponding to the test questions and the completion status, the mastery of knowledge points is quantitatively analyzed and a heat map is built to visualize the mastery. A coordinate matrix is ​​established through the knowledge points corresponding to the test questions, with the horizontal axis representing the knowledge points corresponding to the test questions and the vertical axis representing individual students or student groups. Based on the preset mastery threshold, each student's mastery status for each knowledge point is mapped to a discrete level, including: setting three knowledge point mastery level thresholds of A, B, and C from high to low; and assigning differentiated visual codes to different mastery levels, wherein: areas where the mastery rate of the teaching focus is lower than B or the mastery rate of the related prerequisite knowledge points is lower than C are marked with the first warning color system; areas where the mastery rate of the related prerequisite knowledge points belongs to the interval [C, B] or the mastery rate of the teaching focus belongs to the interval [B, A) are marked with the second transition color system; and areas where the mastery rate of the related prerequisite knowledge points exceeds A or the mastery rate of the teaching focus belongs to the interval (B, 1) are marked with the third saturation color system. The teaching quality is determined by analyzing the proportion of different color-coded areas to the total area. When the first warning color-coded area occupies half of the total area, the teaching quality is judged to be poor. The teaching results are then fed back to the teachers so that they can conduct targeted teaching based on the results.

5. The teaching assistance method based on artificial intelligence according to claim 1, characterized in that: The specific methods for identifying and analyzing teaching focus are as follows: Acquire multidimensional teaching data, and extract key teaching points from different dimensions based on the multidimensional teaching data. The multidimensional teaching data includes course materials, text transcripts of teacher lectures, and classroom interaction data. Extract key teaching points from different dimensions, including: The course materials are assigned initial weights based on the curriculum standards, and the weights are adjusted according to the frequency of cross-chapter references of knowledge points in the materials to obtain the weights of different knowledge points in the course materials, W(Zi), where Zi represents the i-th teaching knowledge point; Based on the audio text of the teacher's lecture, high-frequency keywords that match the knowledge points in the course materials are extracted. The high-frequency keywords are then detected in the speech rate change sections. A weight enhancement coefficient is applied to the knowledge points associated with the speech rate drop areas to obtain the weight U(Zi') of the high-frequency keywords corresponding to the knowledge points in the course materials. Here, Zi' represents the teaching knowledge points in the teacher's lecture audio text that match the teaching knowledge points more than a preset threshold. The high-frequency keywords refer to those that are repeated more than the preset threshold. Based on the frequency and accuracy of question responses in classroom interaction data, a weight K (Zi) is generated for different teaching knowledge points through weighted summation. Initial values ​​are assigned to the teaching knowledge points, and the initial values ​​of the teaching knowledge points are weighted and summed according to the weights of different teaching knowledge points on different dimensions to obtain the importance of different teaching knowledge points. Teaching knowledge points whose importance exceeds a preset threshold are marked as teaching priorities.

6. A teaching assistance system based on artificial intelligence, characterized in that: The teaching support system includes a teaching focus analysis module, a teaching knowledge graph construction module, a test question screening and distribution module, and a teaching quality assessment module. The teaching focus analysis module is used to identify and analyze teaching focus based on multidimensional teaching data. The teaching knowledge graph construction module is used to obtain multi-source teaching datasets through historical teaching data, analyze the dependencies between teaching knowledge points based on the multi-source teaching datasets, and construct a teaching knowledge graph based on the dependencies. The test question filtering and distribution module is used to filter and distribute test questions in the knowledge base based on teaching focus and teaching knowledge graph respectively. The teaching quality assessment module is used to analyze the mastery of teaching focus and related prerequisite knowledge points based on the completion of the distributed test questions, and to analyze, evaluate and provide feedback on the teaching quality based on the mastery level.

7. The teaching assistance system based on artificial intelligence according to claim 6, characterized in that: The teaching focus analysis module includes a dependency direction analysis unit and a dependency degree analysis unit; The dependency direction analysis unit is used to obtain a multi-source teaching dataset through historical teaching data. The multi-source teaching dataset includes a subset of student answer records and a subset of knowledge points. Two knowledge points Za and Zb are arbitrarily selected from the subset of knowledge points, and the dependency direction between the knowledge points is determined based on the timestamps corresponding to the knowledge points Za and Zb. The dependency analysis unit is used to extract the number of questions that correspond to knowledge points Za and Zb from the subset of student answer records when the timestamp of knowledge point Za is earlier than that of knowledge point Zb, and to calculate the co-occurrence frequency of knowledge points Za and Zb based on the number of questions. Extract the correct answer rates of knowledge points Za and Zb, as well as the correct answer rates of the two knowledge points appearing together, from the subset of student answer records. When Y(Za-Zb) is lower than the preset correct answer rate threshold, analyze and judge the correct answer rates of the questions corresponding to the two knowledge points to obtain the degree of dependence of the two knowledge points as [Y(Zb) / Y(Za)].

8. The teaching assistance system based on artificial intelligence according to claim 7, characterized in that: The test question filtering and distribution module includes a question filtering unit and a question distribution unit; The test question filtering unit is used to traverse related prerequisite knowledge points according to the teaching knowledge graph, focusing on the teaching focus and following the dependency direction. It then filters test questions based on matching the knowledge points examined in the knowledge base with the teaching focus and related prerequisite knowledge points. This includes: A test question is randomly selected from the knowledge base. The knowledge point Zn corresponding to the randomly selected test question is represented as the i-th knowledge point. The applicability of the test question is calculated by comparing the knowledge point Zn with the teaching focus and related prerequisite knowledge points. Test questions with applicability exceeding a preset threshold are selected as test questions to be distributed. The test question distribution unit is used to map the degree of interaction between students and teaching focus through multi-dimensional teaching data, classify the degree of interaction, and distribute the test questions corresponding to the interactive teaching focus in reverse order according to their applicability.

9. A teaching assistance system based on artificial intelligence according to claim 8, characterized in that: The teaching quality assessment module includes a heat map construction unit and a teaching quality evaluation unit. The heatmap construction unit is used to quantify the mastery of knowledge points based on the knowledge points corresponding to the test questions and the completion status, and to build a heatmap to visualize the mastery level. A coordinate matrix is ​​established through the knowledge points corresponding to the test questions, with the horizontal axis representing the knowledge points corresponding to the test questions and the vertical axis representing individual students or student groups. Based on the preset mastery threshold, each student's mastery status for each knowledge point is mapped to a discrete level, including: setting three knowledge point mastery level thresholds of A, B, and C from high to low; and assigning differentiated visual codes to different mastery levels, wherein: areas where the mastery rate of the teaching focus is lower than B or the mastery rate of the related prerequisite knowledge points is lower than C are marked with the first warning color system; areas where the mastery rate of the related prerequisite knowledge points belongs to the interval [C, B] or the mastery rate of the teaching focus belongs to the interval [B, A) are marked with the second transition color system; and areas where the mastery rate of the related prerequisite knowledge points exceeds A or the mastery rate of the teaching focus belongs to the interval (B, 1) are marked with the third saturation color system. The teaching quality assessment unit is used to analyze and judge the teaching quality based on the proportion of different color-coded areas to the total area. When the first warning color-coded area occupies half of the total area, the teaching quality is judged to be poor, and the teaching results are fed back so that teachers can conduct targeted teaching based on the results.

Images