Method and system for recommending personalized learning paths for students based on big data analysis
Through big data analysis and student feature evaluation model, combined with teacher suggestions, the learning path is dynamically adjusted, which solves the problem that the existing system fails to fully consider factors, and improves learning efficiency and path matching.
Patent Information
- Application Number
- CN202411246357.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-06
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-09-06
AI Technical Summary
The existing personalized learning path planning system fails to fully consider many factors and the students' own situation, resulting in poor path planning results.
Using a method based on big data analysis, a student feature evaluation model is established by obtaining and preprocessing students' initial data, matching the learning paths of similar historical students, and dynamically adjusting them based on the evaluation scores and teaching resource difficulty levels, combined with teacher correction suggestions, Optimize the learning path.
It realizes dynamic matching and adjustment of learning paths, improves learning efficiency, avoids blind and repetitive labor, and meets the diverse needs of students.
Smart Images

Figure CN118761879B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of big data analysis, and in particular to a method and system for recommending personalized learning paths for students based on big data analysis. Background Art
[0002] With the rapid development of Internet technology, online education, as a new teaching model, has been widely used. Online education breaks the limitations of time and space, allowing students to access educational resources anytime and anywhere, greatly improving the flexibility and convenience of learning.
[0003] However, online education also has some problems. First, the abundance of online education resources often makes it difficult for students to find the most suitable learning content among many courses, which dampens their enthusiasm for learning. Secondly, the traditional one-size-fits-all teaching model cannot meet the personalized needs of students, resulting in unsatisfactory learning efficiency and results. Although personalized learning paths have received much attention in recent years, they have also helped students improve their learning efficiency to a certain extent. However, in the existing systematic research on personalized learning path planning, there is still a lack of full consideration of various factors and planning of learning paths in combination with the students' own circumstances, which leads to unsatisfactory path planning results. Summary of the invention
[0004] (1) Technical issues to be resolved
[0005] The purpose of the present invention is to provide a method and system for recommending personalized learning paths for students based on big data analysis, so as to solve the problem that the learning path is not planned in combination with the students' own situation and various factors are not fully considered, thus resulting in unsatisfactory path planning results.
[0006] (2) Technical solution
[0007] To achieve the above objectives, on the one hand, the present invention provides a method for recommending personalized learning paths for students based on big data analysis, the method comprising:
[0008] S1. Obtaining initial data of the trainee, performing data preprocessing on the initial data to obtain first analysis data, and applying the first analysis data to the trainee characteristic evaluation model to obtain the trainee's teaching attribute data; the initial data includes initial evaluation data and trainee information; the data preprocessing includes data cleaning and data conversion;
[0009] S2. Matching the historical student with the highest similarity according to the teaching attribute data and the student's expectation value and recording the student as a first similar entity, and using the learning path of the first similar entity as the initial recommended learning path for the student;
[0010] S3. Testing the learning outcomes of the students according to a preset teaching progress testing cycle to obtain an evaluation score; performing data analysis based on the evaluation score to obtain a growth curve of the students, and obtaining the growth efficiency rate of the students based on the growth curve as a first contribution rate;
[0011] S4, obtaining recommended teaching resources through correlation analysis of the evaluation scores; matching the history students with the highest similarity between the evaluation scores and the recommended teaching resources and recording them as the second similar body, obtaining the growth benefit rate of the second similar body after learning the teaching resources and recording it as the second contribution rate; calculating the average of the first contribution rate and the second contribution rate to obtain the predicted contribution rate;
[0012] S5. According to the predicted contribution rate and the expected value of the students, the reference learning path is corrected to obtain a first corrected learning path; the predicted contribution rate and the expected value of the students are fed back to the training teacher to obtain teaching correction suggestions, and the first corrected learning path is corrected according to the teaching correction suggestions to obtain a second corrected learning path and recorded as the recommended learning path; when the monitored students update the assessment scores, the recommended learning path is updated; the reference learning path is the recommended learning path obtained from the student's last teaching progress check. If the student has not yet conducted a teaching progress check, the reference learning path is the initial recommended learning path.
[0013] Furthermore, the method of obtaining recommended teaching resources by correlation analysis of the evaluation scores includes:
[0014] The evaluation score is divided according to different preset test items to obtain the evaluation score of each test item and recorded as the project evaluation score, and the test items are sorted according to the project evaluation score to obtain a sorting result;
[0015] Dividing the assessment scores according to knowledge points to obtain knowledge point assessment scores, obtaining the students' historical knowledge point learning data, applying the knowledge point assessment scores and the historical knowledge point learning data to the students' knowledge point mastery degree analysis model to obtain the mastery degree of the knowledge points, obtaining the knowledge points in the teaching resources of the teaching platform system and analyzing them to obtain the difficulty level of the knowledge points; the historical knowledge point learning data includes the learning time and learning frequency of the knowledge points;
[0016] Match teaching resources with corresponding difficulty levels of knowledge points according to the students' mastery of knowledge points and record them as first teaching resources; obtain the correlation between knowledge points in the first teaching resource, and plan the learning path of the first teaching resource according to the correlation and sorting results to obtain recommended teaching resources.
[0017] Furthermore, the teaching resource matching the difficulty level of the corresponding knowledge point according to the student's mastery of the knowledge point is recorded as the first teaching resource; the correlation between the knowledge points in the first teaching resource is obtained, and the method of planning the learning path for the first teaching resource according to the correlation and the ranking result to obtain the recommended teaching resource includes:
[0018] The teaching resources matching the difficulty level of the corresponding knowledge points include intercepting the teaching resources according to the students' mastery of the knowledge points, and the interception is to determine the learning starting point of the teaching resources according to the students' mastery of the knowledge points;
[0019] The sorting result is to sort the test items from low to high according to the evaluation scores; obtain the knowledge points included in the test items, obtain the feature data of the knowledge points, and obtain the correlation between the knowledge points through the association rule mining algorithm; the feature data includes the name of the knowledge point, the difficulty level of the knowledge point, and the description of the knowledge point; when the correlation is less than a preset correlation threshold, the first teaching resource is sorted from low to high according to the evaluation scores of the knowledge points; when the correlation is greater than the preset association threshold, the first teaching resource is sorted according to the correlation; the correlation is the logical order between the knowledge points.
[0020] Furthermore, the teaching resources matching the difficulty level of the corresponding knowledge points include intercepting the teaching resources according to the students' mastery of the knowledge points, and the method of intercepting to determine the learning starting point of the teaching resources according to the students' mastery of the knowledge points includes:
[0021] Acquire historical learning data of students' learning knowledge points, wherein the historical learning data includes error types and error frequencies; derive weak links of knowledge points through data analysis of the historical learning data; trace back to previous knowledge points based on the weak links of the knowledge points, and when the mastery of the previous knowledge points is less than a preset mastery threshold, continue to trace back to the previous knowledge points; when the mastery of the previous knowledge points is greater than the preset mastery threshold, use the previous knowledge points as the learning starting point of the teaching resources.
[0022] Furthermore, the method of obtaining the mastery degree of the knowledge points by using the knowledge point assessment score and the knowledge point historical learning data through the student knowledge point mastery degree analysis model, obtaining the knowledge points in the teaching resources in the teaching platform system and analyzing to obtain the difficulty level of the knowledge points includes:
[0023] Obtain historical learning data and historical assessment scores of history students' learning knowledge points and record them as first feature data, perform feature conversion on the first feature data to obtain second feature data, establish a mapping relationship between the second feature data and the knowledge point mastery degree through a neural network algorithm to obtain a student knowledge point mastery degree analysis model;
[0024] Obtain historical assessment scores of knowledge points learned by history students, calculate the average score of the knowledge points through mean calculation of the historical assessment scores, and convert the assessment scores into a first difficulty coefficient through a linear transformation method; obtain a second difficulty coefficient of the knowledge points through the Delphi method, and perform a linear weighted algorithm on the first difficulty coefficient and the second difficulty coefficient to obtain the difficulty level of the knowledge points.
[0025] Furthermore, the method of obtaining the student's teaching attribute data by using the student characteristic evaluation model through the first analysis data includes:
[0026] The initial assessment data and student information of historical students are obtained and recorded as historical student characteristic data, and the student assessment model is constructed by the historical student characteristic data through a neural network algorithm; the real-time assessment data of historical students is obtained to correct the student assessment model; the teaching attribute data includes the student's learning style data and learning ability data.
[0027] Furthermore, the method of feeding back the predicted contribution rate and the trainee's expected value to the training teacher and obtaining teaching correction suggestions includes:
[0028] The predicted contribution rate and the trainee's expected value are used for data analysis to obtain a gap value, which is then fed back to the training teacher; the training teacher evaluates the first revised learning path based on the gap value to obtain a degree of recognition; when the degree of recognition is greater than a preset degree of recognition threshold, the recommended learning path is executed; when the degree of recognition is less than the preset degree of recognition threshold, the training teacher obtains modification authority and gives teaching correction suggestions; the teaching correction suggestions are preset correction suggestion options.
[0029] Based on the same inventive concept, the present invention also provides a personalized learning path recommendation system for students based on big data analysis, the system comprising: a student teaching attribute acquisition module, an initial recommended learning path acquisition module, a first contribution rate acquisition module, a predicted contribution rate acquisition module, and a recommended learning path acquisition module, and each module is sequentially connected in communication;
[0030] The student teaching attribute acquisition module is used to acquire the student's initial data, obtain the first analysis data after data preprocessing, and obtain the student's teaching attribute data through the student characteristic evaluation model; the initial data includes initial evaluation data and student information; the data preprocessing includes data cleaning and data conversion;
[0031] An initial recommended learning path acquisition module is used to match the historical student with the highest similarity according to the teaching attribute data and the student's expected value and record it as a first similar entity, and use the learning path of the first similar entity as the initial recommended learning path for the student;
[0032] A first contribution rate acquisition module is used to detect the learning outcomes of students according to a preset teaching progress detection cycle to obtain an evaluation score; perform data analysis based on the evaluation score to obtain a growth curve of the students, and obtain the growth efficiency rate of the students according to the growth curve and record it as a first contribution rate;
[0033] A predicted contribution rate acquisition module is used to obtain recommended teaching resources through association analysis of the evaluation scores; match the historical students with the highest similarity between the evaluation scores and the recommended teaching resources and record them as the second similar body, obtain the growth benefit rate of the second similar body after learning the teaching resources and record it as the second contribution rate; calculate the average of the first contribution rate and the second contribution rate to obtain the predicted contribution rate;
[0034] The recommended learning path acquisition module is used to correct the reference learning path according to the predicted contribution rate and the student's expected value to obtain a first corrected learning path; feed back the predicted contribution rate and the student's expected value to the training teacher and obtain teaching correction suggestions, correct the first corrected learning path according to the teaching correction suggestions to obtain a second corrected learning path and record it as the recommended learning path; when the monitored student updates the assessment score, the recommended learning path is updated; the reference learning path is the recommended learning path obtained from the student's last teaching progress check. If the student has not yet conducted a teaching progress check, the reference learning path is the initial recommended learning path.
[0035] (3) Beneficial effects
[0036] Compared with the prior art, the present invention has the following beneficial effects:
[0037] 1. Conduct regular testing and evaluation of students, and dynamically adjust the recommended learning path based on the assessment scores and growth curves to ensure that the learning path always matches the students' current learning status and abilities, thereby enabling students to learn in a more efficient manner and avoiding blind and repetitive work.
[0038] 2. Comprehensive consideration of multiple factors such as students' expectations, assessment scores, difficulty levels of teaching resources, and teachers' teaching correction suggestions makes the recommended learning path more comprehensive and scientific, and can better meet students' learning needs. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 This is a flowchart of a method for recommending personalized learning paths for students based on big data analysis according to Embodiment 1 of the present invention;
[0040] Figure 2 This is a schematic diagram of the module composition of the student personalized learning path recommendation system based on big data analysis in Example 2 of the present invention. DETAILED DESCRIPTION
[0041] 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.
[0042] Before giving examples, it is necessary to explain the application scenarios of the present invention. The present invention is a personalized learning path recommendation method and system for students based on big data analysis. On the teaching platform, it is difficult for many learners to clearly understand their own knowledge level and learning status, and the expected value of learning is constantly adjusted dynamically during the learning process. If the teaching plan is not combined with the expected value of the students, then the students will find it very difficult to learn, and they will feel discouraged and depressed, which may eventually fail to meet expectations. In the case of relatively certain goals, dynamically adjusting teaching resources according to the students' own learning ability and learning attributes will be more conducive to the acquisition of knowledge and rapid growth. At the same time, in order to achieve the expected value of the students, it is necessary to analyze the growth of the students on the knowledge points, and to timely correct the learning path according to the growth, so that the students can learn and gain something, and truly realize differentiated teaching. At the same time, the dynamic information of the students is transmitted to the teaching teacher, and the teaching teacher can combine his own experience to make targeted suggestions and correct guidance, so as to avoid the recommendation of a single system automatic learning path, so as to be more suitable for the growth of the students.
[0043] Example 1: Figure 1 As shown, this embodiment provides a method for recommending personalized learning paths for students based on big data analysis, and the method includes:
[0044] S1. Obtain the initial data of the students, perform data preprocessing on the initial data to obtain the first analysis data, and obtain the teaching attribute data of the students through the student feature evaluation model; the initial data includes initial evaluation data and student information; the data preprocessing includes data cleaning and data conversion; the initial evaluation data is obtained by using a series of standardized tests or questionnaires on the students, and is used to evaluate the basic level of the students in a specific subject or skill field. The evaluation can set the test question type according to the needs, such as multiple-choice questions, fill-in-the-blank questions, short-answer questions and other forms, aiming to fully understand the students' knowledge mastery, learning ability, etc. Collect the basic information of the students and obtain the students' portrait, that is, their teaching attribute data, through the students' evaluation model, so as to better understand their personalized needs. Basic information includes name, age, learning background, learning goals, interests and hobbies, etc. For example, students can be divided into different types, such as openness, responsibility, extroversion, agreeableness and neuroticism. According to the evaluation data and information of the students, the students' knowledge mastery, learning style and learning ability can be obtained. Different test scores indicate differences between individuals.
[0045] S2. Match the historical student with the highest similarity according to the teaching attribute data and the student's expected value and record it as the first similar entity, and use the learning path of the first similar entity as the initial recommended learning path for the student; define multiple similarity metrics for each dimension of the teaching attribute data to compare the similarity between the current student and the historical students, such as the matching degree of learning style, the similarity of knowledge mastery, and the degree of proximity of learning ability, using the defined similarity metrics. This embodiment can convert each dimension of the teaching attribute into a dimension vector, and then use the Euclidean distance calculation method to obtain the similarity. By calculating the similarity between the current student and each historical student in the database, when the obtained similarity value is the smallest, it is the student with the highest similarity, and extract its learning path from the data of the first similar entity as the initial recommended learning path. The initial recommended learning path includes teaching resources and the order and association between them.
[0046] S3. Test the students' learning outcomes according to the preset teaching progress test cycle to obtain the assessment score; perform data analysis based on the assessment score to obtain the students' growth curve, and obtain the students' growth efficiency rate based on the growth curve and record it as the first contribution rate; set regular and irregular teaching progress tests for the students, which can be conducted in the form of test paper evaluation or video question and answer. The evaluation can be used to understand the students' mastery of the knowledge they have learned and their growth. The growth curve is a score change curve drawn through statistical analysis. The slope of the change curve is the students' growth efficiency. The students' learning effect can be obtained through the growth efficiency. When the slope of the change curve is positive, it means that the teaching resources are effective and have achieved certain results. If the slope of the change curve is negative, it means that the teaching resources are arranged unreasonably, which causes students to have certain learning pressure or disorganized learning and no results.
[0047] S4. Obtain recommended teaching resources through correlation analysis of the assessment scores; match the history students with the highest similarity based on the assessment scores and the recommended teaching resources and record them as the second similar entity, obtain the growth benefit rate of the second similar entity after learning the teaching resources and record them as the second contribution rate; calculate the average of the first contribution rate and the second contribution rate to obtain the predicted contribution rate; obtain the students' mastery of the knowledge points based on the assessment scores of the students, obtain teaching resources based on the mastery analysis, match the history students based on the recommended teaching resources and assessment scores, first obtain the history students with the same or closest assessment scores, and then select the history students with the same or closest teaching resources as the similar entity, so that the growth benefit rate of the similar entity after learning the recommended teaching resources can be obtained. The growth benefit rate is the slope value of the growth curve of the similar entity, so that the reference data can be obtained by fully considering the students' own learning situation. The predicted contribution rate can be obtained by calculating the first contribution rate and the second contribution rate of the similar body through the mean. The predicted contribution rate is the student's predicted growth in learning recommended teaching resources. While referring to the similar body, the student's own learning ability can be considered to make a more reasonable prediction. Because learning resources have a certain adaptability, obtaining the learning situation on the reference student has a certain reference value, and the teaching resource recommendation can be made more reasonably. At the same time, the student's own ability will be used as the main reference. Therefore, the mean calculation adopts the weighted mean calculation method. For example, the first contribution rate of the student's own ability is assigned a weight value of 0.8, and the second contribution rate of the similar body is assigned a weight value of 0.2, so that the student's own ability is mainly used. Of course, it is also possible to reverse the weight value based on the statistical analysis method by obtaining the difference between the predicted growth situation and the actual growth situation, and then obtain a statistical weight value.
[0048] S5. According to the predicted contribution rate and the expected value of the students, the reference learning path is corrected to obtain the first corrected learning path; the predicted contribution rate and the expected value of the students are fed back to the training teacher and the teaching correction suggestions are obtained. The first corrected learning path is corrected according to the teaching correction suggestions to obtain the second corrected learning path and recorded as the recommended learning path; when the monitoring students update the evaluation score, the recommended learning path is updated; the reference learning path is the recommended learning path obtained by the student's last teaching progress test. If the student has not yet conducted a teaching progress test, the reference learning path is the initial recommended learning path. The student's expected value is the student's learning goal. At each stage, the student's goal can remain unchanged or be dynamically adjusted. For example, at the beginning, the student's goal is to get an excellent evaluation score, but after a period of study, it is found that the difficulty is high and the student's learning ability cannot effectively keep up with the teaching progress, then the goal may be adjusted to pass. By fully considering the expected value of each student after the test, the teaching resources can be planned more personalized and a learning path that better meets the student's goals can be provided. After the reference learning path is corrected, the predicted contribution rate and expected value are sent by the system to the training teachers. The training teachers have rich teaching experience, which cannot be effectively obtained and evaluated by the system. Therefore, introducing the teaching suggestions of teachers can improve the rationality and efficiency of the students' learning path in a more targeted manner. The training teachers' teaching correction suggestions will be combined with the data fed back by the system, including the students' predicted contribution rate, expected value, and the students' learning situation and teaching resource recommendations recorded by the system. The system will select the teaching correction suggestion options preset by the system, so that the system can further optimize the learning path and generate a personalized learning path for each student.
[0049] Furthermore, the method of obtaining recommended teaching resources by correlation analysis of the evaluation scores includes:
[0050] The evaluation scores are divided according to different preset test items to obtain evaluation scores for each test item and recorded as project evaluation scores. The test items are sorted according to the project evaluation scores to obtain sorting results. The scores obtained from the evaluation include scores of many items. For example, a mathematics test may include multiple-choice questions, fill-in-the-blank questions, application questions and other items, and the project scores are obtained according to different test items.
[0051] The assessment scores are divided according to the knowledge points to obtain the knowledge point assessment scores, the students' knowledge point historical learning data are obtained, the knowledge point assessment scores and the knowledge point historical learning data are used by the students' knowledge point mastery analysis model to obtain the mastery of the knowledge points, the knowledge points in the teaching resources of the teaching platform system are obtained and analyzed to obtain the difficulty level of the knowledge points; the knowledge point historical learning data includes the learning time and learning frequency of the knowledge points; the assessment scores can also be divided according to the tested knowledge points, such as geometric functions, permutations and combinations, integral functions, etc., and the scores of these knowledge points can be counted to obtain the scores of the knowledge points, and then the students' historical learning time and learning frequency of these knowledge points are obtained. For example, if the learning time is long and the learning frequency is high, it means that the degree of mastery is relatively deeper and the knowledge points are understood more thoroughly. Then, combined with the assessment scores, the mastery of the knowledge points can be truly reflected. If the learning time is long and the learning frequency is high, but the assessment score is low, it means that although the knowledge points take more time, the mastery is low, which can also reflect that the students' learning ability is weak. The student knowledge mastery degree analysis model can effectively integrate these factors to conduct analysis, thereby obtaining the student's mastery degree of the knowledge points.
[0052] According to the students' mastery of knowledge points, teaching resources with corresponding knowledge point difficulty levels are matched and recorded as the first teaching resources; the correlation between knowledge points in the first teaching resources is obtained, and the learning path planning of the first teaching resources is performed according to the correlation and the sorting result to obtain recommended teaching resources. The higher the students' mastery of knowledge points, the higher the students' learning ability is, and the learning of knowledge points can enter a stage with a higher difficulty level, that is, teaching resources with high difficulty levels of knowledge points can be recommended to students. Teaching resources with high difficulty levels usually have a relatively short teaching time, and the rhythm of knowledge point explanation is faster, which requires students to have a solid grasp of the basics of knowledge points. The lower the students' mastery of knowledge points, the weaker the students' learning ability is. The learning of knowledge points will recommend students teaching resources with low difficulty levels of knowledge points, so that the teaching time will be longer, and the explanation of knowledge points will focus on details and slowness, so as to ensure that students can clearly understand the knowledge points. After normalizing the knowledge point mastery and the knowledge point difficulty level, the approximate value between the two can be obtained through numerical analysis. The smaller the approximate value, that is, the smaller the difference, the closer the knowledge point mastery and the knowledge point difficulty level are, and the more suitable it is for students to learn. Then teaching resources of different difficulty levels will lead to different learning paths. In the first teaching resources, there are teaching resources of different knowledge points. These teaching resources have certain correlations. For example, generally, you learn mathematics and algebra, elementary geometry, measurement and data first, and then advanced algebra, in-depth geometry, data analysis and probability, etc., or you teach theory first and then practice. According to cognitive laws and learning characteristics, arrange teaching content from simple to complex, from concrete to abstract. At the beginning, the first teaching resources are messy, which is a collection of teaching resources of all knowledge points. According to these correlations and the evaluation scores of the test items, the teaching resources can be used to plan the learning path to obtain recommended teaching resources.
[0053] Furthermore, the teaching resource matching the difficulty level of the corresponding knowledge point according to the student's knowledge point mastery degree is recorded as the first teaching resource; the correlation between the knowledge points in the first teaching resource is obtained, and the method of planning the learning path for the first teaching resource according to the correlation and the sorting result to obtain the recommended teaching resource includes:
[0054] The teaching resources that match the difficulty level of corresponding knowledge points include intercepting the teaching resources according to the students' mastery of the knowledge points, and the interception is to determine the learning starting point of the teaching resources according to the students' mastery of the knowledge points; the teaching resources of the knowledge points obtained are the teaching of the entire knowledge points, so there is no need to repeat the learning of the part of the knowledge point that has been mastered. For example, for geometric functions, if the students' mastery level is judged to be 0.8, it means that the students' mastery is very good, and the basic learning part of the corresponding teaching resources does not need to be learned. It is only necessary to intercept the teaching resources according to the mastery, that is, start learning from where the students need to learn, so as to provide students with learning resources more accurately, avoid students from repeating learning, and waste learning time for students, thereby affecting students' enthusiasm.
[0055] The sorting result is to sort the test items from low to high according to the evaluation score; obtain the knowledge points contained in the test items, obtain the feature data of the knowledge points, and obtain the correlation between the knowledge points through the association rule mining algorithm; the feature data includes the name of the knowledge point, the difficulty level of the knowledge point, and the description of the knowledge point; when the correlation is less than the preset association threshold, the first teaching resource is sorted in the order from low to high according to the evaluation score of the knowledge point; when the correlation is greater than the preset association threshold, the first teaching resource is sorted according to the correlation; the correlation is the logical order between the knowledge points. The lower the evaluation score of the test item, the weaker the student is in this area, and it will be used as a key learning target. The knowledge points contained in each test item can be obtained, and then the teaching resources can be further rationally planned according to the correlation of the knowledge points. For example, the student completed three test items in the mathematics course: Project A (one-variable linear equation), Project B (two-variable linear equation system) and Project C (image and properties of linear function). The evaluation scores of these projects have been obtained and sorted from low to high according to the evaluation scores. The sorting result is: Project A < Project B < Project C. Then analyze the knowledge points and their characteristic data contained in each project: Project A (Linear Equation with One Variable) contains the knowledge point name of Linear Equation with One Variable, the knowledge point difficulty level is elementary, and the knowledge point description is to master the definition, solution and simple application of Linear Equation with One Variable. Project B (System of Linear Equations with Two Variables) contains the knowledge point name of System of Linear Equations with Two Variables, the knowledge point difficulty level is intermediate, and the knowledge point description is to understand the concept of System of Linear Equations with Two Variables, master its solution and application in practical problems. The study of System of Linear Equations with Two Variables is usually based on the mastery of Linear Equations with One Variable, so there is a logical order correlation between them. The knowledge point name of Project C (Graph and Properties of Linear Function) contains the knowledge point name of Linear Function, the knowledge point difficulty level is intermediate, and the knowledge point description is to understand the definition of Linear Function, master its graph drawing method and basic properties. The correlation between these knowledge points is evaluated by association rule mining algorithm. In the above example, the correlation between the system of two-variable linear equations and the linear equation of one variable is stronger because there is a clear logical order between them, that is, learning the linear equation of one variable first and then learning the system of two-variable linear equations. Although the linear function also belongs to the field of algebra, its direct correlation with the system of two-variable linear equations is not as strong as that with the linear equation of one variable. Assume that the correlation obtained by the association rule mining algorithm is that the correlation between the system of two-variable linear equations and the linear equation of one variable is greater than the preset correlation threshold (for example, 0.8), while the correlation between the linear function and these two knowledge points is less than the threshold. Based on the above analysis, the first teaching resource is sorted, giving priority to the logical order (i.e., correlation) of the knowledge points, that is, the teaching resources related to the system of two-variable linear equations will be learned after the teaching resources for learning the linear equation of one variable.For the teaching resources of linear functions, since they are not directly related to the first two knowledge points, the order of teaching resources is arranged according to the evaluation scores.
[0056] Furthermore, the teaching resources matching the difficulty level of the corresponding knowledge points include intercepting the teaching resources according to the students' mastery of the knowledge points, and the method of intercepting to determine the learning starting point of the teaching resources according to the students' mastery of the knowledge points includes:
[0057] Obtain the historical learning data of the students' learning knowledge points, the historical learning data includes error types and error frequencies; use the historical learning data to analyze the weak links of the knowledge points; trace back to the previous knowledge points according to the weak links of the knowledge points, and when the mastery of the previous knowledge points is less than the preset mastery threshold, continue to trace back to the previous knowledge points, and when the mastery of the previous knowledge points is greater than the preset mastery threshold, use the previous knowledge points as the learning starting point of the teaching resources. Collect the historical learning data of each knowledge point from the students' learning records or systems. These historical learning data record the specific types of errors made by students in the process of solving problems or learning, such as calculation errors, concept understanding errors, application errors, etc., and count the number of errors that occur on each knowledge point to evaluate the students' mastery of the knowledge point, so as to use data analysis technology to process and analyze the collected historical learning data and obtain the weak links of the students in the knowledge points. The tracing process is realized by using knowledge graphs, which are based on the logical order and dependency relationship between knowledge. According to the mastery of the previous knowledge points, determine the student's learning starting point. If the mastery is less than the preset mastery threshold, it means that the student has a large learning obstacle on the previous knowledge point, so it is necessary to continue to trace the previous knowledge points. Only by consolidating the foundation of the previous knowledge points again can we ensure that the student can learn, understand and master the difficult knowledge points as quickly as possible. If the mastery is greater than the preset mastery threshold, it means that the student has a good grasp of the previous knowledge point. You can first learn and consolidate this knowledge point and use it as the starting point for learning, and then start learning subsequent knowledge points. According to the determined learning starting point, adjust the student's learning path and provide them with personalized learning resources and guidance to help them more effectively make up for their weaknesses and improve learning effects.
[0058] Furthermore, the method of obtaining the mastery degree of the knowledge points by using the knowledge point assessment score and the knowledge point historical learning data through the student knowledge point mastery degree analysis model, obtaining the knowledge points in the teaching resources in the teaching platform system and analyzing to obtain the difficulty level of the knowledge points includes:
[0059] The historical learning data and historical evaluation scores of the history students' learning knowledge points are obtained and recorded as the first feature data, the first feature data is subjected to feature conversion to obtain the second feature data, and the second feature data is mapped to the knowledge point mastery degree through a neural network algorithm to obtain the student's knowledge point mastery degree analysis model; the historical learning data and historical evaluation scores are collected from the student's learning system, and a suitable neural network structure can be selected according to the complexity of the problem and the characteristics of the data. This embodiment adopts a convolutional neural network (CNN) for model training.
[0060] The historical evaluation scores of the knowledge points learned by history students are obtained, the average scores of the knowledge points are obtained by averaging the historical evaluation scores, and the evaluation scores are converted into the first difficulty coefficient by the linear transformation method; the second difficulty coefficient of the knowledge points is obtained by the Delphi method, and the first difficulty coefficient and the second difficulty coefficient are linearly weighted to obtain the difficulty level of the knowledge points. The average score is obtained by adding all the evaluation scores and then dividing them by the total number of evaluation scores. Linear transformation is a method of scaling and translating a set of values according to a certain ratio. For example, the score range of the evaluation score is [0,100], and the average score is divided by 100, so that the average score is converted to [0,1] to represent the difficulty coefficient. The Delphi method is a method of predicting or evaluating through expert opinions, and a second difficulty coefficient is obtained by combining expert opinions. Because the difficulty judgment of knowledge points has a certain accuracy through the expert evaluation method, it can be used as a consideration of the difficulty coefficient of knowledge points. At the same time, the average score of history students learning knowledge points reflects the average learning ability of most students in the learning process, and can more accurately reflect the difficulty of students' learning knowledge points. For example, the first difficulty coefficient of a certain knowledge point is 0.8, and the assigned weight value is 0.7. The second difficulty coefficient is 0.5, and the assigned weight value is 0.3. The difficulty coefficient obtained by using the linear weighted algorithm is 0.71, which can more comprehensively reflect the difficulty level of the knowledge point.
[0061] Furthermore, the method of obtaining the student's teaching attribute data by using the student characteristic evaluation model through the first analysis data includes:
[0062] The initial assessment data and student information of historical students are obtained and recorded as historical student characteristic data, and the student assessment model is constructed by the historical student characteristic data through a neural network algorithm; the real-time assessment data of historical students is obtained to correct the student assessment model; the teaching attribute data includes the student's learning style data and learning ability data.
[0063] Furthermore, the method of feeding back the predicted contribution rate and the trainee's expected value to the training teacher and obtaining teaching correction suggestions includes:
[0064] The predicted contribution rate and the expected value of the trainee are analyzed by data to obtain a gap value, and the gap value is fed back to the training teacher; the training teacher evaluates the first modified learning path according to the gap value to obtain the recognition degree, and when the recognition degree is greater than the preset recognition threshold, the recommended learning path is executed, and when the recognition degree is less than the preset recognition threshold, the training teacher will obtain the modification authority and give the teaching correction suggestion; the teaching correction suggestion is a preset correction suggestion option. The predicted contribution rate is compared with the expected value of the trainee using a data analysis tool or algorithm to calculate the gap value between the two, which reflects the deviation between the trainee's expectations and the actual prediction results. The gap value is fed back to the training teacher in a clear and understandable way, so that the training teacher can fully understand the mismatch between the trainee's expectations and the model prediction. The training teacher evaluates the recommended learning path according to the gap value and combines his own professional knowledge and teaching experience. The evaluation content includes the applicability of the learning path, the difficulty and the matching degree of the trainee's learning style. The training teacher gives a recognition score or evaluation of the recommended learning path. For example, the scoring items can be preset in the system, and the teacher judges and scores to obtain the recognition. The teacher's approval is compared with the preset approval threshold. If the approval is greater than or equal to the preset approval threshold, the recommended learning path is considered appropriate and the learning path is executed. If the approval is less than the preset approval threshold, the recommended learning path is not directly executed, and the system will allow the training teacher to obtain the permission to modify the recommended learning path. The teacher selects from the preset correction suggestions based on his own judgment and experience. These correction suggestions include adjusting the order of learning content, adding or deleting certain teaching resources, etc. According to the teaching correction suggestions given by the training teacher, the system will automatically adjust the learning path accordingly. The system pushes the final learning path to the trainees, executes the corrected learning path, and continuously monitors the trainees' learning progress and results. Considering the teaching experience of the training teacher can ensure that the recommended learning path not only meets the expectations of the trainees, but also has been professionally recognized and optimized by the training teacher, thereby improving the effectiveness and satisfaction of learning.
[0065] Embodiment 2: Based on the same inventive concept, this embodiment also provides a student personalized learning path recommendation system based on big data analysis, including: a student teaching attribute acquisition module, an initial recommended learning path acquisition module, a first contribution rate acquisition module, a predicted contribution rate acquisition module, and a recommended learning path acquisition module, and each module is sequentially connected in communication;
[0066] The student teaching attribute acquisition module is used to acquire the student's initial data, obtain the first analysis data after data preprocessing, and obtain the student's teaching attribute data through the student characteristic evaluation model; the initial data includes initial evaluation data and student information; the data preprocessing includes data cleaning and data conversion;
[0067] An initial recommended learning path acquisition module is used to match the historical student with the highest similarity according to the teaching attribute data and the student's expected value and record it as a first similar entity, and use the learning path of the first similar entity as the initial recommended learning path for the student;
[0068] A first contribution rate acquisition module is used to detect the learning outcomes of students according to a preset teaching progress detection cycle to obtain an evaluation score; perform data analysis based on the evaluation score to obtain a growth curve of the students, and obtain the growth efficiency rate of the students according to the growth curve and record it as a first contribution rate;
[0069] A predicted contribution rate acquisition module is used to obtain recommended teaching resources through association analysis of the evaluation scores; match the historical students with the highest similarity between the evaluation scores and the recommended teaching resources and record them as the second similar body, obtain the growth benefit rate of the second similar body after learning the teaching resources and record it as the second contribution rate; calculate the average of the first contribution rate and the second contribution rate to obtain the predicted contribution rate;
[0070] The recommended learning path acquisition module is used to correct the reference learning path according to the predicted contribution rate and the student's expected value to obtain a first corrected learning path; feed back the predicted contribution rate and the student's expected value to the training teacher and obtain teaching correction suggestions, correct the first corrected learning path according to the teaching correction suggestions to obtain a second corrected learning path and record it as the recommended learning path; when the monitored student updates the assessment score, the recommended learning path is updated; the reference learning path is the recommended learning path obtained from the student's last teaching progress check. If the student has not yet conducted a teaching progress check, the reference learning path is the initial recommended learning path.
[0071] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0072] Finally, it should be noted that: Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for recommending personalized learning paths for students based on big data analysis. It is characterized in that The method comprises: Acquire the initial data of the trainee, perform data preprocessing on the initial data to obtain first analysis data, and obtain the teaching attribute data of the trainee through the first analysis data through the trainee characteristic evaluation model; the initial data includes initial evaluation data and trainee information; the data preprocessing includes data cleaning and data conversion; According to the teaching attribute data and the expected value of the student, the historical student with the highest similarity is matched and recorded as a first similar entity, and the learning path of the first similar entity is used as the initial recommended learning path for the student; Testing the learning outcomes of students according to a preset teaching progress testing cycle to obtain an evaluation score; performing data analysis based on the evaluation score to obtain a growth curve of the students, and obtaining the growth efficiency rate of the students based on the growth curve and recording it as a first contribution rate; The evaluation scores are analyzed through association to obtain recommended teaching resources; the history student with the highest similarity between the evaluation scores and the recommended teaching resources is matched and recorded as the second similar body, and the growth benefit rate of the second similar body after learning the teaching resources is obtained and recorded as the second contribution rate; the first contribution rate and the second contribution rate are averaged to obtain the predicted contribution rate; According to the predicted contribution rate and the expected value of the trainee, the reference learning path is corrected to obtain a first corrected learning path; the predicted contribution rate and the expected value of the trainee are fed back to the training teacher to obtain teaching correction suggestions, and the first corrected learning path is corrected according to the teaching correction suggestions to obtain a second corrected learning path and recorded as a recommended learning path; when the monitored trainee updates the assessment score, the recommended learning path is updated; the reference learning path is the recommended learning path obtained by the trainee's last teaching progress test. If the trainee has not yet conducted a teaching progress test, the reference learning path is the initial recommended learning path; The method of obtaining recommended teaching resources by correlation analysis of the evaluation scores includes: The evaluation score is divided according to different preset test items to obtain the evaluation score of each test item and recorded as the project evaluation score, and the test items are sorted according to the project evaluation score to obtain a sorting result; Dividing the assessment scores according to knowledge points to obtain knowledge point assessment scores, obtaining the students' historical knowledge point learning data, applying the knowledge point assessment scores and the historical knowledge point learning data to the students' knowledge point mastery degree analysis model to obtain the mastery degree of the knowledge points, obtaining the knowledge points in the teaching resources of the teaching platform system and analyzing them to obtain the difficulty level of the knowledge points; the historical knowledge point learning data includes the learning time and learning frequency of the knowledge points; Match the teaching resources with the corresponding difficulty level of the knowledge points according to the students' mastery of the knowledge points and record them as the first teaching resources; obtain the correlation between the knowledge points in the first teaching resources, and plan the learning path for the first teaching resources according to the correlation and the ranking results to obtain the recommended teaching resources; The method of feeding back the predicted contribution rate and the trainee's expected value to the training teacher and obtaining teaching correction suggestions includes: The predicted contribution rate and the trainee's expected value are used for data analysis to obtain a gap value, which is then fed back to the training teacher; the training teacher evaluates the first revised learning path based on the gap value to obtain a degree of recognition; when the degree of recognition is greater than a preset degree of recognition threshold, the recommended learning path is executed; when the degree of recognition is less than the preset degree of recognition threshold, the training teacher obtains modification authority and gives teaching correction suggestions; the teaching correction suggestions are preset correction suggestion options.
2. According to the method for recommending personalized learning paths for students based on big data analysis according to claim 1, It is characterized in that The method of matching the teaching resources with the corresponding knowledge point difficulty levels according to the students' knowledge point mastery degree and recording them as the first teaching resources; obtaining the correlation between the knowledge points in the first teaching resources, and planning the learning path for the first teaching resources according to the correlation and the ranking result to obtain the recommended teaching resources includes: The teaching resources matching the difficulty level of the corresponding knowledge points include intercepting the teaching resources according to the students' mastery of the knowledge points, and the interception is to determine the learning starting point of the teaching resources according to the students' mastery of the knowledge points; The sorting result is to sort the test items from low to high according to the evaluation scores; obtain the knowledge points included in the test items, obtain the feature data of the knowledge points, and obtain the correlation between the knowledge points through the association rule mining algorithm; the feature data includes the name of the knowledge point, the difficulty level of the knowledge point, and the description of the knowledge point; when the correlation is less than a preset correlation threshold, the first teaching resource is sorted from low to high according to the evaluation scores of the knowledge points; when the correlation is greater than the preset association threshold, the first teaching resource is sorted according to the correlation; the correlation is the logical order between the knowledge points.
3. The method for recommending personalized learning paths for students based on big data analysis according to claim 2, It is characterized in that The teaching resources matching the difficulty level of the corresponding knowledge points include intercepting the teaching resources according to the students' mastery of the knowledge points, and the method of intercepting to determine the learning starting point of the teaching resources according to the students' mastery of the knowledge points includes: Acquire historical learning data of students' learning knowledge points, wherein the historical learning data includes error types and error frequencies; derive weak links of knowledge points through data analysis of the historical learning data; trace back to previous knowledge points based on the weak links of the knowledge points, and when the mastery of the previous knowledge points is less than a preset mastery threshold, continue to trace back to the previous knowledge points; when the mastery of the previous knowledge points is greater than the preset mastery threshold, use the previous knowledge points as the learning starting point of the teaching resources.
4. The method for recommending personalized learning paths for students based on big data analysis according to claim 1, It is characterized in that The method of obtaining the mastery degree of the knowledge points by using the knowledge point assessment score and the knowledge point historical learning data through the student knowledge point mastery degree analysis model, obtaining the knowledge points in the teaching resources in the teaching platform system and analyzing to obtain the difficulty level of the knowledge points includes: Obtain historical learning data and historical assessment scores of history students' learning knowledge points and record them as first feature data, perform feature conversion on the first feature data to obtain second feature data, establish a mapping relationship between the second feature data and the knowledge point mastery degree through a neural network algorithm to obtain a student knowledge point mastery degree analysis model; Obtain historical assessment scores of knowledge points learned by history students, calculate the average score of the knowledge points through mean calculation of the historical assessment scores, and convert the assessment scores into a first difficulty coefficient through a linear transformation method; obtain a second difficulty coefficient of the knowledge points through the Delphi method, and perform a linear weighted algorithm on the first difficulty coefficient and the second difficulty coefficient to obtain the difficulty level of the knowledge points.
5. The method for recommending personalized learning paths for students based on big data analysis according to claim 1, It is characterized in that The method of obtaining the student's teaching attribute data by using the first analysis data through the student characteristic evaluation model includes: The initial assessment data and student information of historical students are obtained and recorded as historical student characteristic data, and the student assessment model is constructed by the historical student characteristic data through a neural network algorithm; the real-time assessment data of historical students is obtained to correct the student assessment model; the teaching attribute data includes the student's learning style data and learning ability data.
6. Personalized learning path recommendation system for students based on big data analysis, It is characterized in that The system includes: a student teaching attribute acquisition module, an initial recommended learning path acquisition module, a first contribution rate acquisition module, a predicted contribution rate acquisition module, and a recommended learning path acquisition module, and each module is sequentially connected to each other; The student teaching attribute acquisition module is used to acquire the student's initial data, obtain the first analysis data after data preprocessing, and obtain the student's teaching attribute data through the student characteristic evaluation model; the initial data includes initial evaluation data and student information; the data preprocessing includes data cleaning and data conversion; An initial recommended learning path acquisition module is used to match the historical student with the highest similarity according to the teaching attribute data and the student's expected value and record it as a first similar entity, and use the learning path of the first similar entity as the initial recommended learning path for the student; A first contribution rate acquisition module is used to detect the learning outcomes of students according to a preset teaching progress detection cycle to obtain an evaluation score; perform data analysis based on the evaluation score to obtain a growth curve of the students, and obtain the growth efficiency rate of the students according to the growth curve and record it as a first contribution rate; A predicted contribution rate acquisition module is used to obtain recommended teaching resources through association analysis of the evaluation scores; match the historical students with the highest similarity between the evaluation scores and the recommended teaching resources and record them as the second similar body, obtain the growth benefit rate of the second similar body after learning the teaching resources and record it as the second contribution rate; calculate the average of the first contribution rate and the second contribution rate to obtain the predicted contribution rate; The recommended learning path acquisition module is used to correct the reference learning path according to the predicted contribution rate and the student's expected value to obtain a first corrected learning path; feed back the predicted contribution rate and the student's expected value to the training teacher and obtain teaching correction suggestions, correct the first corrected learning path according to the teaching correction suggestions to obtain a second corrected learning path and record it as the recommended learning path; when the monitored student updates the assessment score, the recommended learning path is updated; the reference learning path is the recommended learning path obtained from the student's last teaching progress check. If the student has not yet conducted a teaching progress check, the reference learning path is the initial recommended learning path.
