Artificial intelligence-based comprehensive data simulation teaching evaluation method and system

Abstract

The invention discloses an artificial intelligence-based comprehensive data simulation teaching evaluation method and system, and particularly relates to the technical field of teaching evaluation, and the method comprises the steps: generating a knowledge graph of teaching resources based on an artificial intelligence technology, obtaining the knowledge point progress information of students based on the positions and relevance of knowledge points related to exercises in the knowledge graph, and carrying out the teaching evaluation. Knowledge points of the knowledge graph are clustered, knowledge point difference information of students is determined, a learning result evaluation coefficient is generated, and learning duration information and learning habit information of the students are determined by analyzing use data of the students on a learning platform based on records of using learning resources by the students in the learning platform; according to the method, the usual learning performance evaluation coefficient is generated, the learning progress and the learning attitude of the student are evaluated by analyzing the performance of the student in exercises and the performance of using the learning platform, the method is helpful for the teacher to examine the learning performance of the student, and more accurate support is provided for the teacher to make a teaching decision.

Description

Technical Field

[0001] The present invention relates to the technical field of teaching assessment, and more specifically, to a comprehensive data simulation teaching assessment method and system based on artificial intelligence. Background Art

[0002] With the development of educational informatization, the digitization of teaching resources and the popularization of online learning platforms, how to conduct comprehensive evaluation and personalized tutoring of students' learning through data analysis and intelligent technology has become an important issue in the field of modern education. Existing technologies can monitor students' learning activities on university learning platforms, but student evaluation methods still rely on regular test scores, which often cannot fully reflect students' learning process and learning attitude.

[0003] In order to solve the above defects, a technical solution is now provided. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a comprehensive data simulation teaching and evaluation method and system based on artificial intelligence to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A comprehensive data simulation teaching and evaluation method based on artificial intelligence, specifically comprising the following steps:

[0007] S1: Generate a knowledge graph of teaching resources based on artificial intelligence technology according to the teaching resources uploaded by the course teachers and the historical teaching resources in the learning platform;

[0008] S2: Based on the position and relevance of the knowledge points involved in the exercises in the knowledge graph, obtain the students' knowledge point progress information, and determine the students' knowledge point difference information by clustering the knowledge points in the knowledge graph to generate the learning outcome evaluation coefficient;

[0009] S3: Based on the records of students using learning resources on the learning platform, by analyzing the students' usage data on the learning platform, determine the students' learning time information and learning habits information, and generate the usual learning performance evaluation coefficient;

[0010] S4: Evaluate students’ learning progress and learning attitude by analyzing their performance on exercises and their use of the learning platform.

[0011] In a preferred embodiment, the knowledge point progress information includes:

[0012] The knowledge point progress information is represented by the knowledge level coefficient;

[0013] The logic for obtaining the knowledge level coefficient is as follows: according to the historical records of students using exercises, based on the positions and relevance of the knowledge points involved in the exercises in the knowledge graph, determine the one or more knowledge points corresponding to each exercise on the knowledge graph, and determine the student's proficiency coefficient on the knowledge point according to the number of answers and correctness of the student on the specific knowledge point;

[0014] Calculate the student's proficiency coefficient on the knowledge point. The calculation formula is: SL n =ZQ n ln(DT n +1); among them, SL n is the student's proficiency coefficient on the knowledge point, ZQ n is the correct rate of students regarding the nth knowledge point, DT n is the number of times the student answered questions about the nth knowledge point, n = 1, 2, 3, ..., N, N is a positive integer, and n is the number of knowledge points in the knowledge graph;

[0015] The importance coefficients of different knowledge points are determined according to the dependency relationship of the knowledge graph. In the knowledge graph, each knowledge point is regarded as a node, and the relationship between knowledge points constitutes an edge. The importance coefficient of each knowledge point is determined by the PageRank algorithm. The expression is: ; Among them, PR n The importance coefficient of each knowledge point, H(n) is the set of all predecessor nodes pointing to node N, L i is the out-degree of node i, indicating the number of links sent from node i, PR i is the importance coefficient of node i, d is the damping factor;

[0016] Calculate the knowledge level coefficient, the calculation formula is: Among them, SP zs is the knowledge level coefficient.

[0017] In a preferred embodiment, the knowledge point difference information includes:

[0018] The knowledge point difference information is represented by the knowledge point clustering deviation coefficient;

[0019] The logic for obtaining the knowledge point clustering deviation coefficient is: by using the K-means clustering algorithm on the knowledge graph, based on the distance between each node in the knowledge graph, determine the cluster to which different nodes in the knowledge graph belong, including:

[0020] Initialize K initial cluster centers. For each knowledge point, assign the knowledge point to the closest cluster according to the distance between the knowledge point and each cluster center. The distance is defined as: d(Q m ,Q j)=length(path(m,j)),d(Q m ,Q j ) is the minimum path between node m and node j;

[0021] For each cluster, calculate the sum of the path distances from each node in the cluster to other nodes, select the node with the smallest total path distance as the new cluster center, and compare the cluster allocation of the current iteration with that of the previous iteration. If the cluster allocation has not changed, the algorithm converges and the iteration is terminated. If the cluster allocation has changed, return to step 2 and continue iterating.

[0022] According to the historical records of students' exercises, the exercises that students answered correctly are extracted, and the knowledge points for answering the correct exercises are obtained. The number of knowledge points for answering the correct exercises in each cluster is counted, and the number of knowledge points for answering the correct exercises in each cluster is marked as: LX k , where k = 1, 2, 3, ..., K, K is a positive integer, and k is the number of the cluster;

[0023] Get the mean and standard deviation of the number of knowledge points in each cluster regarding the correct answers to the exercises, and label the mean and standard deviation of the number of knowledge points in each cluster regarding the correct answers to the exercises as: LX avg and LX std ,in,

[0024] Calculate the knowledge point clustering deviation coefficient, the calculation formula is: Among them, JL pc is the knowledge point clustering deviation coefficient.

[0025] In a preferred embodiment, the learning outcome assessment coefficient includes

[0026] Through comprehensive analysis of knowledge point progress information and knowledge point difference information, the knowledge level coefficient and knowledge point clustering deviation coefficient are weighted and calculated to build a learning outcome evaluation model and generate a learning outcome evaluation coefficient. The calculation formula of the learning outcome evaluation coefficient is: Among them, pg cg is the learning outcome evaluation coefficient, α1 and α2 are the proportional coefficients of the knowledge level coefficient and the knowledge point clustering deviation coefficient, respectively, and α1>α2>0.

[0027] In a preferred embodiment, the learning duration information includes:

[0028] The learning time information is represented by the effective learning coefficient;

[0029] The logic for obtaining the effective learning coefficient is as follows: according to the student's learning activity record in the learning platform, determine the duration of each time the student uses the learning platform, the number of times the learning resources are used each time, and the length of time the student is unresponsive in the learning platform each time, and mark the duration of each time the student uses the learning platform, the number of times the learning resources are used each time, and the length of time the student is unresponsive in the learning platform each time as: SC f , CS f TL f ; Wherein, f = 1, 2, 3, ..., F, F is a positive integer, and f is the number of each time a student logs into the learning platform;

[0030] Calculate the effective learning coefficient, the calculation formula is: Among them, YX xx is the effective learning coefficient.

[0031] In a preferred embodiment, the learning habit information includes:

[0032] The learning habit information is expressed by the habit difference variation coefficient;

[0033] The logic for obtaining the habit difference variation coefficient is as follows: collect the usage interval data of students on the learning platform, calculate the time difference between two adjacent learning sessions based on each login time of the students, and mark the time difference between the two adjacent learning sessions as: T c , where c = 1, 2, 3, ..., C, C is a positive integer, and c is the number of two consecutive learnings;

[0034] By using the autoregressive model to predict the students' learning interval, the students' usage interval data on the learning platform are divided into a training group and a control group, and the time difference between two adjacent learning in the training group is marked as: T a , the time difference between two consecutive learning in the control group is marked as: T b , where a=1, 2, 3, ..., A, b=1, 2, 3, ..., B, B and A are positive integers respectively, and a+b=c;

[0035] The expression of the autoregressive model is: y =φ1T y-1 +φ2T y-2 +......+φ p T y-p +ε y ; Among them, T y is the predicted time difference between two consecutive learnings, T y-1 , T y-2 ,......,T y-p is the time difference data between two consecutive learnings in the training group, φ1, φ2, ..., φ pis the coefficient of the autoregressive model, ε y is the noise term;

[0036] The autoregressive model is trained using the data in the training group to predict the data in the control group, and the predicted data in the control group is marked as: YC b ;

[0037] Calculate the coefficient of variation of habit difference, the calculation formula is: Among them, BH cy is the coefficient of variation of habit difference.

[0038] In a preferred embodiment, the usual learning performance evaluation coefficient includes:

[0039] By comprehensively analyzing the students' study time information and study habit information, weighted calculation of the effective learning coefficient and the habit difference change coefficient, a daily study performance evaluation model is constructed to generate the daily study performance evaluation coefficient, which is expressed as: Among them, pg bx is the evaluation coefficient of usual learning performance, β1 and β2 are the proportional coefficients of effective learning coefficient and habit difference change coefficient, respectively, β1>β2>0.

[0040] In a preferred embodiment, the analysis of the student's performance in the exercises and the performance in using the learning platform includes:

[0041] By analyzing the students' performance in the exercises and their performance in using the learning platform, the learning outcome evaluation coefficient and the usual learning performance evaluation coefficient are weighted and summed to determine the student's assessment score. The calculation formula for the assessment score is: PF kh =λ1pg cg +λ2pg bx PF kh is the student's assessment score, λ1 and λ2 are the weight factors of the learning outcome evaluation coefficient and the usual learning performance evaluation coefficient, respectively, and λ1 and λ2 are both greater than 0;

[0042] Set the assessment score threshold, compare the student's assessment score with the assessment score threshold, and generate a pass signal if the assessment score is greater than the assessment score threshold; if the assessment score is less than the assessment score threshold, generate a fail signal.

[0043] In a preferred embodiment, a comprehensive data simulation teaching and evaluation system based on artificial intelligence includes a graph construction module, a data acquisition module, a learning achievement evaluation module, a daily learning performance evaluation module and an assessment and scoring module, and the modules are signal-connected;

[0044] The graph construction module is used to generate and manage the knowledge graph of teaching resources through artificial intelligence technology. Based on the teaching resources uploaded by the course teachers and the historical teaching resources already in the learning platform, it automatically constructs the graph structure involving the relationship and position of each knowledge point;

[0045] A data collection module, used to represent the knowledge point progress information through the knowledge level coefficient, represent the knowledge point difference information through the knowledge point clustering deviation coefficient, represent the learning time information through the effective learning coefficient, and represent the learning habit information through the habit difference change coefficient;

[0046] The learning achievement evaluation module is used to perform weighted calculation on the knowledge level coefficient and the knowledge point clustering deviation coefficient, construct a learning achievement evaluation model, and generate a learning achievement evaluation coefficient;

[0047] The daily learning performance evaluation module is used to perform weighted calculation on the effective learning coefficient and the habit difference change coefficient, build a daily learning performance evaluation model, and generate a daily learning performance evaluation coefficient;

[0048] The assessment and scoring module is used to perform weighted summation of the learning outcome assessment coefficient and the usual learning performance assessment coefficient to determine the student's assessment score, and generate different signals based on the student's assessment score.

[0049] Technical effects and advantages of the present invention:

[0050] The present invention can understand students' mastery of various knowledge points through analysis based on knowledge graphs, and can dynamically track students' learning progress by analyzing information such as students' performance in exercises, knowledge point progress and learning time. The present invention helps to accurately evaluate students' daily learning situation, helps teachers to assess students' learning performance, and gives students more reasonable daily grades, thereby avoiding the possibility that students' final grades may be affected by poor performance in just one exam. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0052] Figure 1 It is a flow chart of a comprehensive data simulation teaching and evaluation method based on artificial intelligence of the present invention;

[0053] Figure 2 The present invention is a structural diagram of an artificial intelligence-based comprehensive data simulation teaching and evaluation system. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Example 1

[0056] like Figure 1 A flowchart of a comprehensive data simulation teaching and evaluation method based on artificial intelligence is given, which specifically includes the following steps:

[0057] S1: Generate a knowledge graph of teaching resources based on artificial intelligence technology according to the teaching resources uploaded by the course teachers and the historical teaching resources in the learning platform;

[0058] S2: Based on the position and relevance of the knowledge points involved in the exercises in the knowledge graph, obtain the students' knowledge point progress information, and determine the students' knowledge point difference information by clustering the knowledge points in the knowledge graph to generate the learning outcome evaluation coefficient;

[0059] S3: Based on the records of students using learning resources on the learning platform, by analyzing the students' usage data on the learning platform, determine the students' learning time information and learning habits information, and generate the usual learning performance evaluation coefficient;

[0060] S4: Evaluate students’ learning progress and learning attitude by analyzing their performance on exercises and their use of the learning platform.

[0061] In step 1, in college learning, teachers upload teaching resources to the learning platform. Teaching resources include teaching videos, textbooks, courseware, exercises, review materials, etc., to ensure that students can check them at any time after class, so that students can review as needed and consolidate their knowledge. Students can log in to the learning platform through identity authentication, and flexibly arrange their learning progress on the platform according to their personal learning rhythm, watch course videos, complete after-class exercises, download materials, etc., to deepen their understanding of knowledge points;

[0062] Teachers can view students' activities on the learning platform and give them more reasonable grades based on their test scores, thus preventing students from having their final grades affected by a single poor performance on a test.

[0063] Among them, according to the teaching resources uploaded by the course teachers and the historical teaching resources in the learning platform, a knowledge graph of teaching resources is generated based on artificial intelligence technology, including knowledge extraction, representation, relationship modeling and dynamic updating of the graph.

[0064] Natural Language Processing (NLP): The core of knowledge graph is to extract knowledge points from text data (such as books, academic papers, online resources, etc.). NLP technology extracts key information from text through processing steps such as word segmentation, part-of-speech tagging, entity recognition, and relationship extraction. In particular, named entity recognition (NER) technology can identify entities in knowledge graphs (such as names, place names, terms, etc.), laying the foundation for knowledge organization. After identifying knowledge points (entities), AI identifies the relationship between these knowledge points through relationship extraction models;

[0065] Based on Graph Neural Network (GNN): In the knowledge graph, knowledge points and their relationships can be represented as a graph structure. Graph neural network can operate on the graph structure through deep learning methods to discover hidden relationships or infer new relationships. GNN can learn the complex interactions between knowledge points, thereby automatically discovering potential relationships and expanding the knowledge graph;

[0066] Knowledge completion based on relational reasoning: In knowledge graphs, incomplete knowledge is often encountered. AI can use rule reasoning (such as Markov Logi c Networks, MLN) or deep learning reasoning (such as TransE model) to reason about existing knowledge, infer possible relationships, and thus complete the graph.

[0067] In step 2, based on the records of students using learning resources in the learning platform, the students' learning activities are mapped to the knowledge graph, the students' learning progress and mastery in the knowledge graph are determined, and exercises on the corresponding knowledge points are provided to students. According to the students' usual use of exercises, the students' knowledge point progress information and knowledge point difference information are obtained, the knowledge point progress information is represented by the knowledge level coefficient, and the knowledge point difference information is represented by the knowledge point clustering deviation coefficient.

[0068] The logic for obtaining the knowledge level coefficient is as follows: according to the historical records of students using exercises, based on the positions and relevance of the knowledge points involved in the exercises in the knowledge graph, determine the one or more knowledge points corresponding to each exercise on the knowledge graph, and determine the student's proficiency coefficient on the knowledge point according to the number of answers and correctness of the student on the specific knowledge point;

[0069] For example: a student has done 100 questions in total, and 10 of them involve mathematical function knowledge points. Therefore, the number of times the student answered questions about mathematical function knowledge points is 10. The student answered the questions involving mathematical function knowledge points correctly 4 times in total, so the student's accuracy rate about mathematical function knowledge points is 0.4.

[0070] Calculate the student's proficiency coefficient on the knowledge point. The calculation formula is: SL n =ZQ nln(DT n +1); among them, SL n is the student's proficiency coefficient on the knowledge point, ZQ n is the correct rate of students regarding the nth knowledge point, DT n is the number of times the student answered questions about the nth knowledge point, n = 1, 2, 3, ..., N, N is a positive integer, and n is the number of knowledge points in the knowledge graph;

[0071] The importance coefficients of different knowledge points are determined according to the dependency relationship of the knowledge graph. In the knowledge graph, each knowledge point is regarded as a node, and the relationship between knowledge points constitutes an edge. The importance coefficient of each knowledge point is determined by the PageRank algorithm. The expression is: Among them, PR n The importance coefficient of each knowledge point, H(n) is the set of all predecessor nodes pointing to node N, L i is the out-degree of node i, indicating the number of links sent from node i, PR i is the importance coefficient of node i, d is the damping factor, usually 0.85;

[0072] It should be noted that a large importance coefficient usually means that these knowledge points have more predecessor or successor knowledge points. They may be basic or core in the learning process, and therefore have a great impact on other knowledge points. Knowledge points with a small importance coefficient may be more specialized or advanced knowledge points, or they may not be knowledge points that every student needs to master in the learning process. These knowledge points may play a supporting role in the entire learning path and depend on other basic knowledge points. Therefore, when students have a high proficiency in knowledge points with a large importance coefficient, it means that students have significant learning outcomes in core knowledge and have mastered the important knowledge points in the learning path very well.

[0073] Calculate the knowledge level coefficient, the calculation formula is: Among them, SP zs is the knowledge level coefficient.

[0074] From the formula, we can see that the larger the knowledge level coefficient is, the higher the degree of mastery of the knowledge points can be determined based on the students' use of exercises, and different knowledge points may have a high degree of mastery.

[0075] The logic for obtaining the knowledge point clustering deviation coefficient is: by using the K-means clustering algorithm on the knowledge graph, based on the distance between each node in the knowledge graph, determine the cluster to which different nodes in the knowledge graph belong, including:

[0076] It should be noted that the distance between each node in the knowledge graph is determined by automatically learning the similarity between knowledge points through the graph neural network in artificial intelligence. For example, it is determined based on the relationship between nodes in the knowledge graph (subordinate relationship, causal relationship, dependency relationship, etc.), including Euclidean distance, cosine similarity, etc.

[0077] Initialize K initial cluster centers. For each knowledge point, assign the knowledge point to the closest cluster according to the distance between the knowledge point and each cluster center. The distance is defined as: d(Q m ,Q j )=length(path(m,j)),d(Q m ,Q j ) is the minimum path between node m and node j;

[0078] For each cluster, calculate the sum of the path distances from each node in the cluster to other nodes, select the node with the smallest total path distance as the new cluster center, and compare the cluster allocation of the current iteration with that of the previous iteration. If the cluster allocation has not changed, the algorithm converges and the iteration is terminated. If the cluster allocation has changed, return to step 2 and continue iterating.

[0079] According to the historical records of students' exercises, the exercises that students answered correctly are extracted, and the knowledge points for answering the correct exercises are obtained. The number of knowledge points for answering the correct exercises in each cluster is counted, and the number of knowledge points for answering the correct exercises in each cluster is marked as: LX k , where k = 1, 2, 3, ..., K, K is a positive integer, and k is the number of the cluster;

[0080] Get the mean and standard deviation of the number of knowledge points in each cluster regarding the correct answers to the exercises, and label the mean and standard deviation of the number of knowledge points in each cluster regarding the correct answers to the exercises as: LX avg and LX std ,in,

[0081] Calculate the knowledge point clustering deviation coefficient, the calculation formula is: Among them, JL pc is the knowledge point clustering deviation coefficient.

[0082] It can be seen from the formula that the larger the knowledge point clustering deviation coefficient is, the more inconsistent the students' mastery of knowledge points in the clusters divided by the knowledge graph using the clustering algorithm is. The difference in knowledge points mastered by students in each cluster may be large, indicating that students may have a bias in certain subjects, that is, they perform better in some knowledge areas and weaker in other areas.

[0083] Through comprehensive analysis of knowledge point progress information and knowledge point difference information, the knowledge level coefficient and knowledge point clustering deviation coefficient are weighted and calculated to build a learning outcome evaluation model and generate a learning outcome evaluation coefficient. The calculation formula of the learning outcome evaluation coefficient is: Among them, pg cg is the learning outcome evaluation coefficient, α1 and α2 are the proportional coefficients of the knowledge level coefficient and the knowledge point clustering deviation coefficient, respectively, and α1>α2>0.

[0084] From the formula, we can see that the larger the knowledge level coefficient and the smaller the knowledge point clustering deviation coefficient, the larger the learning outcome evaluation coefficient, which means that the students may have mastered more knowledge points and their mastery of the knowledge points may be relatively even. Conversely, the smaller the knowledge level coefficient and the larger the knowledge point clustering deviation coefficient, the smaller the learning outcome evaluation coefficient, which means that the students may have mastered fewer knowledge points and their mastery of the knowledge points may be uneven.

[0085] In step 3, based on the records of students using learning resources on the learning platform, the students' learning time information and learning habit information are collected, and the learning time information is represented by the effective learning coefficient, and the learning habit information is represented by the habit difference change coefficient.

[0086] It should be noted that the learning time information reflects the time students spend focusing on learning activities. Through the learning time information, we can observe the students' enthusiasm for learning, which can help teachers make auxiliary decisions in the evaluation of students. By incorporating behavioral data such as learning time into the evaluation system, teachers can evaluate students' learning status more fairly and comprehensively.

[0087] Learning habit information reflects the time series characteristics of students' learning behavior through the autoregressive model, which helps to evaluate the regularity and changing trends of students' learning behavior. If students' learning behavior can be well predicted by past learning behavior with a small error, it means that students' learning habits are relatively stable and regular. If students' learning behavior has large fluctuations or cannot be effectively predicted by historical data, it means that students' learning habits are unstable and there may be frequent learning intervals or sudden learning patterns. This can help teachers evaluate students' learning attitudes more comprehensively.

[0088] The logic for obtaining the effective learning coefficient is as follows: according to the student's learning activity record in the learning platform, determine the duration of each time the student uses the learning platform, the number of times the learning resources are used each time, and the length of time the student is unresponsive in the learning platform each time, and mark the duration of each time the student uses the learning platform, the number of times the learning resources are used each time, and the length of time the student is unresponsive in the learning platform each time as: SC f , CS f TL f; Wherein, f = 1, 2, 3, ..., F, F is a positive integer, and f is the number of each time a student logs into the learning platform;

[0089] It should be noted that the duration of each time a student uses the learning platform represents the total duration from the beginning of the student's identity authentication when entering the learning platform to the student's exit from the learning platform; the number of times the learning resources are used each time represents the number of times the student clicks on the learning resources on the learning platform, that is, the number of times the student views teaching videos, exercises, review materials and other learning resources on the learning platform; the duration of each unresponsiveness on the learning platform represents the length of time the student is unresponsive on the learning platform.

[0090] Calculate the effective learning coefficient, the calculation formula is: Among them, YX xx is the effective learning coefficient.

[0091] From the formula, we can see that represents the square root of the time the f-th student spent on the learning platform. By using the square root, we can avoid the excessive impact of extreme learning time on the overall score. ln(CS f +1) represents the natural logarithm of the number of times the f-th student uses learning resources on the learning platform. By amplifying the impact of the number of times the resources are used on the overall score through the natural logarithm, it can reflect the diversity of students' learning and avoid the excessive impact of extreme values ​​on the results. The larger the effective learning coefficient, the higher the student's enthusiasm in the learning platform and the better the evaluation result for the student.

[0092] The logic for obtaining the habit difference variation coefficient is as follows: collect the usage interval data of students on the learning platform, calculate the time difference between two adjacent learning sessions based on each login time of the students, and mark the time difference between the two adjacent learning sessions as: T c , where c = 1, 2, 3, ..., C, C is a positive integer, and c is the number of two consecutive learnings;

[0093] By using the autoregressive model to predict the students' learning interval, the students' usage interval data on the learning platform are divided into a training group and a control group, and the time difference between two adjacent learning in the training group is marked as: T a , the time difference between two consecutive learning in the control group is marked as: T b , where a=1, 2, 3, ..., A, b=1, 2, 3, ..., B, B and A are positive integers respectively, and a+b=c;

[0094] The expression of the autoregressive model is: y =φ1T y-1 +φ2T y-2 +......+φ p T y-p +εy ; Among them, T y is the predicted time difference between two consecutive learnings, T y-1 , T y-2 ,......,T y-p is the time difference data between two consecutive learnings in the training group, φ1, φ2, ..., φ p is the coefficient of the autoregressive model, ε y is the noise term;

[0095] The autoregressive model is trained using the data in the training group to predict the data in the control group, and the predicted data in the control group is marked as: YC b ;

[0096] Calculate the coefficient of variation of habit difference, the calculation formula is: Among them, BH cy is the coefficient of variation of habit difference.

[0097] From the formula, we can see that the larger the coefficient of variation of habit differences is, the more likely it is that students’ learning behaviors may be irregular, or that there are sudden learning behavior patterns.

[0098] By comprehensively analyzing the students' study time information and study habit information, weighted calculation of the effective learning coefficient and the habit difference change coefficient, a daily study performance evaluation model is constructed to generate the daily study performance evaluation coefficient, which is expressed as: Among them, pg bx is the evaluation coefficient of usual learning performance, β1 and β2 are the proportional coefficients of effective learning coefficient and habit difference change coefficient, respectively, β1>β2>0.

[0099] From the formula, we can see that the larger the effective learning coefficient is and the smaller the habit difference change coefficient is, the larger the usual learning performance evaluation coefficient is, which means that the student's usual learning attitude is more serious, and the teacher can give the student a higher score in the evaluation.

[0100] In step 4, by analyzing the student's performance in the exercises and the performance in using the learning platform, the learning outcome evaluation coefficient and the usual learning performance evaluation coefficient are weighted and summed to determine the student's assessment score. The calculation formula for the assessment score is: PF kh =λ1pg cg +λ2pg bx PF kh is the student's assessment score, λ1 and λ2 are the weight factors of the learning outcome evaluation coefficient and the usual learning performance evaluation coefficient, respectively, and λ1 and λ2 are both greater than 0.

[0101] Set the assessment score threshold and compare the student's assessment score with the assessment score threshold. If the assessment score is greater than the assessment score threshold, a pass signal is generated, indicating that the student's performance on the learning platform is good and a higher score can be provided for the student's usual performance. If the assessment score is less than the assessment score threshold, a fail signal is generated, indicating that the student's performance on the learning platform is poor. The teacher can judge whether there is a "poor performance in the exam" by comparing the student's test scores with the platform learning performance.

[0102] The present invention can understand students' mastery of various knowledge points through analysis based on knowledge graphs, and can dynamically track students' learning progress by analyzing information such as students' performance in exercises, knowledge point progress and learning time. The present invention can accurately evaluate students' daily learning situation, help teachers assess students' learning performance, give students more reasonable daily grades, and avoid students' final grades being affected by poor performance in just one exam.

[0103] Example 2

[0104] like Figure 2 A structural diagram of a comprehensive data simulation teaching and evaluation system based on artificial intelligence is given, which specifically includes a graph construction module, a data acquisition module, a learning outcome evaluation module, a daily learning performance evaluation module, and an assessment and scoring module, and the signal connections between the modules;

[0105] The graph construction module is used to generate and manage the knowledge graph of teaching resources through artificial intelligence technology. Based on the teaching resources uploaded by the course teachers and the historical teaching resources already in the learning platform, it automatically constructs the graph structure involving the relationship and position of each knowledge point;

[0106] A data collection module, used to represent the knowledge point progress information through the knowledge level coefficient, represent the knowledge point difference information through the knowledge point clustering deviation coefficient, represent the learning time information through the effective learning coefficient, and represent the learning habit information through the habit difference change coefficient;

[0107] The learning achievement evaluation module is used to perform weighted calculation on the knowledge level coefficient and the knowledge point clustering deviation coefficient, construct a learning achievement evaluation model, and generate a learning achievement evaluation coefficient;

[0108] The daily learning performance evaluation module is used to perform weighted calculation on the effective learning coefficient and the habit difference change coefficient, build a daily learning performance evaluation model, and generate a daily learning performance evaluation coefficient;

[0109] The assessment and scoring module is used to perform weighted summation of the learning outcome assessment coefficient and the usual learning performance assessment coefficient to determine the student's assessment score, and generate different signals based on the student's assessment score.

[0110] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0111] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0112] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0113] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0114] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0115] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0116] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0117] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A comprehensive data simulation teaching and evaluation method based on artificial intelligence, characterized in that: The specific steps include: S1: Generate a knowledge graph of teaching resources based on artificial intelligence technology according to the teaching resources uploaded by the course teachers and the historical teaching resources in the learning platform; S2: Based on the position and relevance of the knowledge points involved in the exercises in the knowledge graph, obtain the students' knowledge point progress information, and determine the students' knowledge point difference information by clustering the knowledge points in the knowledge graph to generate the learning outcome evaluation coefficient; S3: Based on the records of students using learning resources on the learning platform, by analyzing the students' usage data on the learning platform, determine the students' learning time information and learning habits information, and generate the usual learning performance evaluation coefficient; S4: Evaluate students’ learning progress and learning attitude by analyzing their performance on exercises and their use of the learning platform.

2. According to the comprehensive data simulation teaching and evaluation method based on artificial intelligence in claim 1, it is characterized in that: Knowledge point progress information, including: The knowledge point progress information is represented by the knowledge level coefficient; The logic for obtaining the knowledge level coefficient is as follows: according to the historical records of students using exercises, based on the positions and relevance of the knowledge points involved in the exercises in the knowledge graph, determine the one or more knowledge points corresponding to each exercise on the knowledge graph, and determine the student's proficiency coefficient on the knowledge point according to the number of answers and correctness of the student on the specific knowledge point; Calculate the student's proficiency coefficient on the knowledge point. The calculation formula is: SL n =ZQ n ln(DT n +1); among them, SL n is the student's proficiency coefficient on the knowledge point, ZQ n is the correct rate of students regarding the nth knowledge point, DT n is the number of times the student answered questions about the nth knowledge point, n = 1, 2, 3, ..., N, N is a positive integer, and n is the number of knowledge points in the knowledge graph; The importance coefficients of different knowledge points are determined according to the dependency relationship of the knowledge graph. In the knowledge graph, each knowledge point is regarded as a node, and the relationship between knowledge points constitutes an edge. The importance coefficient of each knowledge point is determined by the PageRank algorithm. The expression is: Among them, PR n The importance coefficient of each knowledge point, H(n) is the set of all predecessor nodes pointing to node N, L i is the out-degree of node i, indicating the number of links sent from node i, PR i is the importance coefficient of node i, d is the damping factor; Calculate the knowledge level coefficient, the calculation formula is: Among them, SP zs is the knowledge level coefficient.

3. The comprehensive data simulation teaching and evaluation method based on artificial intelligence according to claim 2 is characterized in that: Knowledge point difference information, including: The knowledge point difference information is represented by the knowledge point clustering deviation coefficient; The logic for obtaining the knowledge point clustering deviation coefficient is: by using the K-means clustering algorithm on the knowledge graph, based on the distance between each node in the knowledge graph, determine the cluster to which different nodes in the knowledge graph belong, including: Initialize K initial cluster centers. For each knowledge point, assign the knowledge point to the closest cluster according to the distance between the knowledge point and each cluster center. The distance is defined as: d(Q m ,Q j )=length(path(m,j)),d(Q m ,Q j ) is the minimum path between node m and node j; For each cluster, calculate the sum of the path distances from each node in the cluster to other nodes, select the node with the smallest total path distance as the new cluster center, and compare the cluster allocation of the current iteration with that of the previous iteration. If the cluster allocation has not changed, the algorithm converges and the iteration is terminated. If the cluster allocation has changed, return to step 2 and continue iterating. According to the historical records of students' exercises, the exercises that students answered correctly are extracted, and the knowledge points for answering the correct exercises are obtained. The number of knowledge points for answering the correct exercises in each cluster is counted, and the number of knowledge points for answering the correct exercises in each cluster is marked as: LX k , where k = 1, 2, 3, ..., K, K is a positive integer, and k is the number of the cluster; Get the mean and standard deviation of the number of knowledge points in each cluster regarding the correct answers to the exercises, and label the mean and standard deviation of the number of knowledge points in each cluster regarding the correct answers to the exercises as: LX avg and LX std ,in, Calculate the knowledge point clustering deviation coefficient, the calculation formula is: Among them, JL pc is the knowledge point clustering deviation coefficient.

4. The comprehensive data simulation teaching and evaluation method based on artificial intelligence according to claim 3 is characterized in that: Learning outcomes assessment factors, including Through comprehensive analysis of knowledge point progress information and knowledge point difference information, the knowledge level coefficient and knowledge point clustering deviation coefficient are weighted and calculated to build a learning outcome evaluation model and generate a learning outcome evaluation coefficient. The calculation formula of the learning outcome evaluation coefficient is: Among them, pg cg is the learning outcome evaluation coefficient, α1 and α2 are the proportional coefficients of the knowledge level coefficient and the knowledge point clustering deviation coefficient, respectively, and α1>α2>0.

5. According to the comprehensive data simulation teaching and evaluation method based on artificial intelligence according to claim 4, it is characterized in that: Study duration information, including: The learning time information is represented by the effective learning coefficient; The logic for obtaining the effective learning coefficient is as follows: according to the student's learning activity record in the learning platform, determine the duration of each time the student uses the learning platform, the number of times the learning resources are used each time, and the length of time the student is unresponsive in the learning platform each time, and mark the duration of each time the student uses the learning platform, the number of times the learning resources are used each time, and the length of time the student is unresponsive in the learning platform each time as: SC f , CS f TL f ; Wherein, f = 1, 2, 3, ..., F, F is a positive integer, and f is the number of each time a student logs into the learning platform; Calculate the effective learning coefficient, the calculation formula is: Among them, YX xx is the effective learning coefficient.

6. The method for comprehensive data simulation teaching and evaluation based on artificial intelligence according to claim 5 is characterized in that: Study habits information, including: The learning habit information is expressed by the habit difference variation coefficient; The logic for obtaining the habit difference variation coefficient is as follows: collect the usage interval data of students on the learning platform, calculate the time difference between two adjacent learning sessions based on each login time of the students, and mark the time difference between the two adjacent learning sessions as: T c , where c = 1, 2, 3, ..., C, C is a positive integer, and c is the number of two consecutive learnings; By using the autoregressive model to predict the students' learning interval, the students' usage interval data on the learning platform are divided into a training group and a control group, and the time difference between two adjacent learning in the training group is marked as: T a , the time difference between two consecutive learning in the control group is marked as: T b , where a=1, 2, 3, ..., A, b=1, 2, 3, ..., B, B and A are positive integers respectively, and a+b=c; The expression of the autoregressive model is: y =φ1T y-1 +φ2T y-2 +……+φ p T y-p +ε y ; Among them, T y is the predicted time difference between two consecutive learnings, T y-1 , T y-2 ,……,T y-p is the time difference data between two consecutive learnings in the training group, φ1, φ2, ..., φ p is the coefficient of the autoregressive model, ε y is the noise term; The autoregressive model is trained using the data in the training group to predict the data in the control group, and the predicted data in the control group is marked as: YC b ; Calculate the coefficient of variation of habit difference, the calculation formula is: Among them, BH cy is the coefficient of variation of habit difference.

7. The method of comprehensive data simulation teaching and evaluation based on artificial intelligence according to claim 6 is characterized in that: Evaluation coefficients of daily learning performance include: By comprehensively analyzing the students' study time information and study habit information, weighted calculation of the effective learning coefficient and the habit difference change coefficient, a daily study performance evaluation model is constructed to generate the daily study performance evaluation coefficient, which is expressed as: Among them, pg bx is the evaluation coefficient of usual learning performance, β1 and β2 are the proportional coefficients of effective learning coefficient and habit difference change coefficient, respectively, β1>β2>0.

8. The comprehensive data simulation teaching and evaluation method based on artificial intelligence according to claim 7 is characterized in that: Analyze students’ performance on exercises and on the learning platform, including: By analyzing the students' performance in the exercises and their performance in using the learning platform, the learning outcome evaluation coefficient and the usual learning performance evaluation coefficient are weighted and summed to determine the student's assessment score. The calculation formula for the assessment score is: PF kh =λ1pg cg +λ2pg bx PF kh is the student's assessment score, λ1 and λ2 are the weight factors of the learning outcome evaluation coefficient and the usual learning performance evaluation coefficient, respectively, and λ1 and λ2 are both greater than 0; Set the assessment score threshold, compare the student's assessment score with the assessment score threshold, and generate a pass signal if the assessment score is greater than the assessment score threshold; if the assessment score is less than the assessment score threshold, generate a fail signal.

9. An artificial intelligence-based comprehensive data simulation teaching and evaluation system, used to implement an artificial intelligence-based comprehensive data simulation teaching and evaluation method according to any one of claims 1 to 8, characterized in that: It includes a graph construction module, a data collection module, a learning outcome assessment module, a daily learning performance assessment module, and an assessment and scoring module, and the signals between the modules are connected; The graph construction module is used to generate and manage the knowledge graph of teaching resources through artificial intelligence technology. Based on the teaching resources uploaded by the course teachers and the historical teaching resources already in the learning platform, it automatically constructs the graph structure involving the relationship and position of each knowledge point; A data collection module, used to represent the knowledge point progress information through the knowledge level coefficient, represent the knowledge point difference information through the knowledge point clustering deviation coefficient, represent the learning time information through the effective learning coefficient, and represent the learning habit information through the habit difference change coefficient; The learning achievement evaluation module is used to perform weighted calculation on the knowledge level coefficient and the knowledge point clustering deviation coefficient, construct a learning achievement evaluation model, and generate a learning achievement evaluation coefficient; The daily learning performance evaluation module is used to perform weighted calculation on the effective learning coefficient and the habit difference change coefficient, build a daily learning performance evaluation model, and generate a daily learning performance evaluation coefficient; The assessment and scoring module is used to perform weighted summation of the learning outcome assessment coefficient and the usual learning performance assessment coefficient to determine the student's assessment score, and generate different signals based on the student's assessment score.

Images