Progressive planning effect data-driven self-learning and self-improvement weighting evaluation method and system, electronic equipment and storage medium

Abstract

The invention relates to the technical field of learning planning, and discloses a progressive planning effect data-driven self-learning and self-improvement weighting evaluation method and system, electronic equipment and a storage medium, and the method comprises the steps: analyzing the difficulty of knowledge points, and carrying out the layering of the difficulty according to the pre-repairing relation and difficulty change rate between the knowledge points; and performing statistics on learning behaviors of the learner to analyze the mastering state of the knowledge points, and performing weighting on the knowledge points based on the learning duration and the mastering state. According to the method, the learning paths of all the knowledge points are determined through the pre-repairing relationship, the knowledge units are constructed, and the average difficulty of each unit is calculated, so that the difficulty of the knowledge points in the learning process can be effectively identified, and the learning efficiency is improved by performing fine-grained monitoring on each learning behavior of a learner. According to the method, the fine change of the learner in the learning process can be understood more deeply, the repeated learning weight of the knowledge points is adjusted through the learning duration management value of the learner on the knowledge points, and the time distribution of the learner on each knowledge point can be effectively balanced.

Description

Technical Field

[0001] The present invention relates to the field of learning planning technology, and specifically to a self-learning and self-improvement weighted evaluation method, system, electronic device and storage medium driven by progressive planning effect data. Background Art

[0002] With the popularization of learning tablets and other devices, self-learning learning methods are gradually accepted and used by students and parents. With the development of information technology and the widespread application of online education platforms, personalized learning has become one of the important directions of current educational research. The existing personalized learning path planning methods mainly rely on simple statistics and analysis of learners' past learning records;

[0003] However, in actual applications, learners' mastery status and learning behaviors are highly dynamic and individual. This static planning method is difficult to adapt to learners' personalized learning needs, resulting in a mismatch between the learning path and the learners' actual situation. In the difficulty analysis of knowledge points, most existing methods only consider a single difficulty indicator of the knowledge point, ignoring the prerequisite relationship between knowledge points and the dynamic changes in learners' mastery of different knowledge points. This leads to unclear difficulty stratification and discontinuous learning process in learning path planning, resulting in a low overall effect of learning path planning. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is that the learner's mastery status and learning behavior are highly dynamic and individual. This static planning method is difficult to adapt to the learner's personalized learning needs, resulting in a mismatch between the learning path and the learner's actual situation. In the difficulty analysis of knowledge points, most of the existing methods only consider a single difficulty index of the knowledge point, ignoring the prerequisite relationship between knowledge points and the dynamic changes in the learner's mastery of different knowledge points, resulting in unclear difficulty stratification and discontinuous learning process in the learning path planning, resulting in a low overall effect of the learning path planning.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a self-learning and self-improvement weighted evaluation method driven by progressive planning effect data, comprising:

[0007] Obtain learners' self-study results and process parameters, analyze the difficulty of knowledge points, and stratify the difficulty according to the prerequisite relationship and difficulty change rate between knowledge points;

[0008] Statistics on learners’ self-study behavior are used to analyze the mastery of knowledge points. Based on the learning time and mastery status, weights are assigned to knowledge points to analyze the comprehensive priority of knowledge points and allocate learning time.

[0009] Define the objective function, use the multi-objective optimization algorithm to plan the learning path, build the fusion population through non-dominated sorting, and iteratively select the optimal solution of the learning path;

[0010] Regularly calculate learners' learning indicators, collect learners' learning behavior data and mastery status to monitor learning progress.

[0011] As a preferred solution of the self-learning and self-improvement weighted evaluation method driven by progressive planning effect data of the present invention, wherein: the self-learning results include: conducting self-learning tests on self-learning students based on the self-learning question bank of the jth knowledge point, and calculating the average score of the students on the jth knowledge point;

[0012] The process parameters include building a knowledge point self-study question bank and a self-study video bank based on self-study knowledge points, building a self-study group chat based on social software, and recording the number of students' discussions;

[0013] Based on the self-study video library of the jth knowledge point, count the average number of views of students;

[0014] Based on the jth knowledge point, count the average number of times students discuss the jth knowledge point;

[0015] The difficulty of the knowledge point is calculated by the average test score, average number of views, and average number of discussions, expressed as:

[0016] ;

[0017] in, represents the difficulty of knowledge point j, represents the average score of the self-study test for knowledge point j, represents the average number of views of the self-study video library of knowledge point j, represents the average number of discussions on knowledge point j, , and They represent the lowest average score among all knowledge points, the highest average number of views among all knowledge points, and the highest average number of comments among all knowledge points respectively;

[0018] Through the logical association of textbook materials and knowledge points, determine the knowledge points that must be learned before learning all knowledge points, and clarify the prerequisite relationship between knowledge points;

[0019] Based on the prerequisite relationship of knowledge points, knowledge point j and all the prerequisite knowledge points are constructed into knowledge units. , and calculate the knowledge unit The average difficulty of all knowledge points ;

[0020] According to the average difficulty Sort all knowledge units by size and calculate the average difficulty of all knowledge units The difficulty change rate between , and calculate the mean and standard deviation of all difficulty change rates. If the average difficulty of the knowledge unit If the difficulty change rate between them is greater than or equal to the sum of the mean and double standard deviation of all difficulty change rates, then it is judged that there is a difficulty gap;

[0021] according to According to the size of the knowledge units, the knowledge points are sorted and the difficulty levels are stratified according to the difficulty gaps.

[0022] As a preferred solution of the self-learning and self-improvement weighted evaluation method driven by progressive planning effect data of the present invention, wherein: the statistical learner's self-learning learning behavior analysis of the mastery status of knowledge points includes calculating the video learning completion rate based on the video viewing time and total time of the learner i's self-learning video library for knowledge point j;

[0023] Count the total learning time of learner i in learning knowledge point j, including the viewing time of the self-study video library, the reading time and the note-taking time;

[0024] Count the number of discussions that learner i has on learning knowledge point j;

[0025] According to the statistical data of learner i, including the video learning completion rate, the total learning time, the maximum video learning completion rate, learning time and discussion times of all learners, normalization is performed;

[0026] The normalized video learning completeness rate, total learning time and the sum of discussion times are used as the learner's learning behavior data, and normalized according to the maximum learning behavior value of all learners i to obtain the learning behavior value of learner i for knowledge point j;

[0027] Count the learner's abnormal behavior of watching videos in the self-study video library when learning knowledge point j, and count the number of non-learning related discussions when the learner is learning knowledge point j. Sum and normalize the abnormal behavior and non-learning related discussions;

[0028] Based on the actual learning time of learner i for learning knowledge point j divided by the planned learning time, the learning time management value is calculated and normalized;

[0029] The learner's mastery of knowledge points is calculated based on the learning behavior value, abnormal learning behavior value, learning time management value and self-study test average score, expressed as:

[0030] ;

[0031] in, represents the mastery status value of learner i on knowledge point j, represents the self-study test score of learner i on knowledge point j, represents the learning behavior value of learner i in the learning process of knowledge point j, represents the abnormal learning behavior value of learner i learning knowledge point j, It represents the learning time management value of learner i for learning knowledge point j;

[0032] Count the historical average mastery status of all learners on all knowledge points, and calculate the overall standard deviation of the mastery status of all learners;

[0033] If the learner's mastery status value of a knowledge point is greater than or equal to the sum of the historical average mastery status and the overall standard deviation of the mastery status, then the learner is judged to have a good grasp of the knowledge point; if the learner's mastery status value of a knowledge point is less than the sum of the historical average mastery status and the overall standard deviation of the mastery status, then the learner is judged to have a poor grasp of the knowledge point.

[0034] As a preferred solution of the self-learning and self-improvement weighted evaluation method driven by progressive planning effect data of the present invention, wherein: the weighting of knowledge points based on learning time and mastery status, analysis of the comprehensive priority of knowledge points and allocation of learning time include, according to the average value of the learner's learning time management value for different knowledge points, calculating the repeated learning weight of knowledge point j, expressed as:

[0035] ;

[0036] in, represents the weight of knowledge point j, represents the average learning time management value of learner i;

[0037] Taking into account the empowerment weight, the difficulty of knowledge point j and the mastery status value, the comprehensive priority of the knowledge point is calculated and expressed as:

[0038] ;

[0039] in, Represented as the comprehensive priority of knowledge point j, It is expressed as the difficulty value of knowledge point j, represents the set of all previous knowledge points of knowledge point j, Indicates the mastery status value of learner i on the previous knowledge point u;

[0040] Taking into account the priority of knowledge point j and the total planned learning time, the target learning time of the knowledge point is allocated, which is expressed as:

[0041] ;

[0042] in, represents the total learning time, M represents the total number of knowledge points, Represents the allocated learning time for knowledge point j.

[0043] As a preferred solution of the self-learning and self-improvement weighted evaluation method driven by progressive planning effect data of the present invention, wherein: the definition of the objective function and the use of a multi-objective optimization algorithm to plan the learning path include using a multi-objective optimization algorithm NSGA-II to dynamically plan the learner's learning path, initialize the population C and the population individual h, the individual h represents the learning path and the corresponding learning time allocation plan;

[0044] For individuals in the population, according to the comprehensive priority of knowledge points and allocate time The learning path sequence and learning time classification sequence are defined respectively, which are expressed as:

[0045] ;

[0046] ;

[0047] in, represents a sequence of learning paths, represents the learning time allocation sequence, represents the Vth knowledge point, represents the Vth time allocation value;

[0048] According to the learning cycle t in multiple learning, the objective functions of maximizing the learning effect and minimizing the rationality of time allocation are defined respectively, which are expressed as:

[0049] ;

[0050] ;

[0051] in represents the objective function of maximizing learning effect, represents the objective function of minimizing the rationality of time allocation, Represents the total number of knowledge points, and They represent the mastery status of learner i on knowledge point j in learning cycle t and learning cycle t-1 respectively, represents the learning priority of knowledge point k, represents the learning time allocated to knowledge point j in learning cycle t;

[0052] Define the dominance relationship of individuals in a population, including the individuals in the population Less than or equal to individual , then it means that the individual Dominate , count the number of times individuals in the population are dominated;

[0053] If individual x has no dominant individuals, then the individual is defined as non-dominated. Let the set of all non-dominated individuals be ;

[0054] The dominance relationship is determined for each pair of individuals in the population, and the individuals in the population are constructed into an initial dominated set. The individuals with a dominated number of 0 are recorded as the first non-dominated set, and the individuals are removed from the population. The dominated number of the individuals in the dominated set is reduced by 1. If the dominated number of the individuals in the dominated set is reduced to 0, they are placed in the non-dominated set of the next layer. The process is repeated to place the individuals in different non-dominated sets to obtain a non-dominated sorting of the individuals.

[0055] As a preferred solution of the self-learning and self-improvement weighted evaluation method driven by progressive planning effect data of the present invention, wherein: the construction of the fusion population by non-dominated sorting and the iterative selection of the optimal solution of the learning path include: based on the individuals with low non-dominated layers in the non-dominated sorting, as the parent individuals, the learning path intersection and the learning time intersection are performed on the learning path sequence and the time allocation sequence to generate the child individuals;

[0056] Construct parent individuals and offspring individuals into a fusion population , and repeat the non-dominated sorting;

[0057] The current fusion population The individuals in the first non-dominated layer in are taken as the optimal solution set of the Pareto solution set, and the learning effect objective function is selected The solution with the maximum value is taken as the optimal solution of the learning path;

[0058] Apply the selected optimal solution for the learning path to the learner's next learning cycle.

[0059] As a preferred solution of the self-learning and self-improvement weighted evaluation method driven by progressive planning effect data of the present invention, the periodic calculation of the learner's learning indicators includes testing the learner's learning progress and periodically calculating the learner's learning indicators, which is expressed as:

[0060] ;

[0061] in represents the learning index of learner i, M represents the total number of knowledge points, represents the self-study test score of learner i on knowledge point j, represents the highest score of knowledge point j;

[0062] The average learning index of all learners is used as the learning threshold. If the learner's learning index is lower than the learning threshold, it is judged as lower than the learning expectation.

[0063] The collecting of the learner's learning behavior data and mastery status to monitor the learning progress includes, based on the determination that the learning result is lower than the learning expectation, further collecting the learner's learning behavior data and mastery status to monitor the learning progress, including the allocation and completion of learning time, the mastery status of knowledge points and the self-study test scores;

[0064] The time allocation and completion status, knowledge point mastery status and self-study test scores of all learners are counted separately, and based on the average of time allocation and completion status, knowledge point mastery status and self-study test scores, further self-learning evaluation is conducted on learners whose learning performance is below expectations. If any two of the time allocation and completion status, knowledge point mastery status and self-study test scores are below the average, the self-learning level is evaluated to be poor, and the path planning for the next learning cycle is re-performed.

[0065] A self-learning and self-improving weighted evaluation system driven by progressive planning effect data using any method described in the present invention, characterized in that:

[0066] The stratification module obtains learners' self-study results and process parameters, analyzes the difficulty of knowledge points, and stratifies the difficulty according to the prerequisite relationship and difficulty change rate between knowledge points;

[0067] The statistical module counts learners’ self-study behavior to analyze the mastery of knowledge points, assigns weights to knowledge points based on learning time and mastery status, analyzes the comprehensive priority of knowledge points and allocates learning time;

[0068] The optimization module defines the objective function, uses a multi-objective optimization algorithm to plan the learning path, constructs a fusion population through non-dominated sorting, and iteratively selects the optimal solution for the learning path;

[0069] The monitoring module regularly calculates learners’ learning indicators, collects learners’ learning behavior data and mastery status to monitor learning progress.

[0070] A computer device comprises: a memory and a processor; the memory stores a computer program, wherein: the processor implements the steps of any one of the methods of the present invention when executing the computer program.

[0071] A computer-readable storage medium stores a computer program, wherein: when the computer program is executed by a processor, the steps of any one of the methods of the present invention are implemented.

[0072] Beneficial effects of the present invention: The self-learning and self-improvement empowerment evaluation method driven by progressive planning effect data provided by the present invention determines the learning path of all knowledge points through the prerequisite relationship, constructs knowledge units, and calculates the average difficulty of each unit, which can effectively identify the difficulty of knowledge points in the learning process, and can more deeply understand the subtle changes of learners in the learning process by fine-grained monitoring of each learning behavior of the learners. The repeated learning weight of the knowledge points is adjusted by the learners' learning time management value of the knowledge points, which can effectively balance the learners' time allocation to each knowledge point. The learners' learning path is dynamically planned through the NSGA-II multi-objective optimization algorithm, and the maximization of learning effect and the minimization of the rationality of time allocation are used as the optimization objective function, thereby realizing the comprehensive optimization of the learning path, and dynamically adjusting the learning path and learning time allocation plan according to factors such as the learners' mastery status, learning time allocation and knowledge point priority in different learning cycles. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0074] Figure 1 The overall flow chart of the progressive planning effect data-driven self-learning and self-improvement weighted evaluation method provided in the first embodiment of the present invention;

[0075] Figure 2 A schematic diagram of the learning path planning process in the self-learning and self-improvement weighted evaluation method driven by progressive planning effect data provided in the first embodiment of the present invention. DETAILED DESCRIPTION

[0076] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0077] Example 1, reference Figure 1 and Figure 2 , which is an embodiment of the present invention, provides a self-learning and self-improvement weighted evaluation method driven by progressive planning effect data, including:

[0078] S1, analyze the difficulty of knowledge points and stratify the difficulty according to the prerequisite relationship and difficulty change rate between knowledge points.

[0079] Preferably, the difficulty of the knowledge points is analyzed, and the difficulty is stratified according to the prerequisite relationship and difficulty change rate between the knowledge points, including: building a knowledge point self-study question bank and a self-study video bank based on the self-study knowledge points, building a self-study group chat based on social software, and recording the number of students' discussions; based on the self-study question bank of the jth knowledge point, conducting a self-study test on the self-study students, and calculating the students' average score for the jth knowledge point; based on the self-study video bank of the jth knowledge point, counting the average number of students' views; based on the jth knowledge point, counting the average number of students' discussions on the jth knowledge point; and comprehensively calculating the knowledge point difficulty through the average score, average number of views and average number of discussions of the test, expressed as:

[0080] ;

[0081] in represents the difficulty of knowledge point j, represents the average score of the self-study test for knowledge point j, represents the average number of views of the self-study video library of knowledge point j, represents the average number of discussions on knowledge point j, , and They represent the lowest average score among all knowledge points, the highest average number of views among all knowledge points, and the highest average number of comments among all knowledge points respectively.

[0082] The formula for calculating the difficulty of knowledge points is , and Normalization processing has been performed so that students' mastery of knowledge points can be comprehensively reflected through test scores, the number of times they self-study through videos, and the number of discussions, and the impact of changes in mastery on the overall difficulty can be more accurate, so as not to mask or amplify the effects of other behavioral items. After the above corrections and normalization processing, the difficulty value of each knowledge point will more reasonably reflect students' mastery of the knowledge points, learning behavior and interaction, and will not lead to misjudgment of difficulty due to the extreme value of a certain indicator.

[0083] Through the logical association of textbook materials and knowledge points, determine the knowledge points that must be learned before learning all knowledge points, and clarify the prerequisite relationship between knowledge points. Based on the prerequisite relationship of knowledge points, knowledge point j and all prerequisite knowledge points are constructed into knowledge units , and calculate the knowledge unit The average difficulty of all knowledge points .

[0084] According to the average difficulty Sort all knowledge units by size and calculate the average difficulty of all knowledge units The difficulty change rate between , and calculate the mean and standard deviation of all difficulty change rates. If the average difficulty of the knowledge unit If the difficulty change rate between the two is greater than or equal to the sum of the mean and double standard deviation of all difficulty change rates, then it is judged that there is a difficulty gap. The knowledge points are sorted by size and the difficulty is stratified.

[0085] The difficulty of knowledge points is calculated comprehensively through the three dimensions of average test score, average number of views and average number of discussions, and these dimensions are normalized to effectively eliminate the interference of extreme data, making the difficulty assessment of knowledge points more accurate. The three dimensions of test scores, number of views and number of discussions of self-study videos are combined to not only reflect students' performance in the test, but also take into account other behavioral items in the learning process, and more comprehensively reflect students' actual mastery of knowledge points. The learning path of all knowledge points is determined through the prerequisite relationship, knowledge units are constructed, and the average difficulty of each unit is calculated. This can effectively identify the "fault" of knowledge point difficulty in the learning process, that is, the situation where the difficulty of some knowledge points is obviously higher or lower than that of other knowledge points, thereby avoiding the situation where students have too many difficulties in the learning path regulation. Sudden jumps in difficulty in the learning plan ensure the continuity and scientific nature of the learning path. By calculating the difficulty change rate of all knowledge units and using the mean and double standard deviation as thresholds, the existence of difficulty gaps can be determined. Changes in the difficulty of knowledge points can be dynamically monitored, and the learning path can be discovered and adjusted in a timely manner so that learners can gradually improve during the learning process rather than suddenly encountering knowledge points with increased difficulty, thereby reducing learning pressure and improving learning outcomes. Social software is used to record the number of students' discussions, and combined with self-study test scores and video viewing times, students' learning behavior data can be deeply mined. Through these data, students' learning habits, knowledge point mastering preferences, etc. can be accurately evaluated, providing strong support for subsequent intelligent learning path planning and personalized teaching.

[0086] S2, statistics on learners’ self-study behavior to analyze the mastery status of knowledge points, assign weights to knowledge points based on learning time and mastery status, analyze the comprehensive priority of knowledge points and allocate learning time.

[0087] Preferably, statistics are collected on the learner's self-study learning behavior to analyze the mastery status of the knowledge points, including calculating the video learning completion rate based on the video viewing time and total time of the self-study video library of learner i for knowledge point j; counting the total learning time of learner i in learning knowledge point j, including the viewing time of watching the self-study video library, the learning time of reading materials and taking notes; counting the number of discussions of learner i in learning knowledge point j; performing normalization processing based on the statistical data of learner i, including the video learning completion rate, the total learning time, the maximum video learning completion rate for all learners, the learning time and the number of discussions; taking the sum of the normalized video learning completion rate, the total learning time and the number of discussions as the learner's learning behavior data, and performing normalization processing based on the maximum learning behavior value of all learners i, and calculating the learning behavior value of learner i for knowledge point j.

[0088] The abnormal behaviors of learners in watching videos in the self-study video library when learning knowledge point j are counted, including abnormal behaviors such as skipping, fast-forwarding, and early exiting when watching videos. At the same time, the number of non-learning related discussions discussed by learners when learning knowledge point j is counted. The number of abnormal behaviors and non-learning related discussions is summed and normalized, which is expressed as:

[0089] ;

[0090] in represents the abnormal learning behavior value of learner i learning knowledge point j, represents the maximum learning abnormal behavior value of knowledge point j, Indicates the standardized assignment of the kth abnormal behavior, where 0 indicates no abnormality and 1 indicates an abnormality. represents the kth learning behavior value of learner i in the process of learning knowledge point j, and N represents the number of types of learning behavior values;

[0091] Based on the actual learning time of learner i for learning knowledge point j divided by the planned learning time, the learning time management value is calculated and normalized;

[0092] The learner's mastery of knowledge points is calculated based on the learning behavior value, abnormal learning behavior value, learning time management value and self-study test average score, expressed as:

[0093] ;

[0094] in represents the mastery status value of learner i on knowledge point j, represents the self-study test score of learner i on knowledge point j, represents the learning behavior value of learner i in the learning process of knowledge point j, represents the abnormal learning behavior value of learner i learning knowledge point j, Represents the learning time management value of learner i learning knowledge point j.

[0095] Statistics are collected for all learners on all knowledge points in terms of their historical average mastery status, and the overall standard deviation of all learners’ mastery status is calculated. If the learner’s mastery status value for a knowledge point is greater than or equal to the sum of the historical average mastery status and the overall standard deviation of the mastery status, then the learner is judged to have a good grasp of the knowledge point. If the learner’s mastery status value for a knowledge point is less than the sum of the historical average mastery status and the overall standard deviation of the mastery status, then the learner is judged to have a poor grasp of the knowledge point.

[0096] By conducting fine-grained monitoring of each learning behavior of learners (such as video learning completion rate, learning time and number of discussions), and combining it with abnormal learning behaviors (such as skipping, fast-forwarding, and early exiting of video viewing), we can have a deeper understanding of the subtle changes in learners in the learning process. By comprehensively calculating the learning behavior value, abnormal learning behavior value, learning time management value and self-study test score, we can scientifically quantify the learners' mastery of knowledge points, and count the learners' skipping, fast-forwarding, early exiting and other behaviors during video viewing, as well as the number of non-learning related discussions. This can effectively identify the learners' abnormal behaviors in the learning process. According to the learners' learning behavior value, abnormal learning behavior value, learning time management value and self-study test score and other information, we can dynamically adjust the learning path and strategy, prioritize the review of the knowledge point or provide additional learning resources to help learners consolidate and master the weak links.

[0097] In the formula for calculating the learner's mastery of knowledge points, a logistic regression model is used to calculate the learner's mastery of a certain knowledge point j. It is expressed as a comprehensive function of multiple influencing factors, and the result is normalized to the interval [0,1] using the Sigmoid function, which is convenient for describing the learner's mastery of the knowledge point. All parameters are normalized to ensure that the relative influence of each parameter is consistent, avoiding calculation deviations caused by a parameter that is too large or too small in value. In this way, multiple aspects such as learners' learning performance, behavioral habits, and time management ability can be comprehensively considered to describe the learner's overall mastery of a certain knowledge point.

[0098] Furthermore, the knowledge points are weighted based on the learning time and mastery status, and the comprehensive priority of the knowledge points and the allocation of learning time are analyzed, including the calculation of the mastery status value of the learner i for the knowledge point j, considering the influence of the learner's mastery status on the knowledge point weight, and calculating the repeated learning weight of the knowledge point j according to the average value of the learner's learning time management value for different knowledge points, which is expressed as:

[0099] ;

[0100] in represents the weight of knowledge point j, represents the average learning time management value of learner i; The calculation of represents the degree of inadequacy of learner i’s mastery of knowledge point j. The lower the mastery, the higher the weight. The calculation is used to smooth the impact of time management, avoid excessive weight when time management is extremely low, and reflect the learner's time management ability. When the learner's average learning time management value is low, the weight will increase, prompting the learner to allocate more time to the study of this knowledge point. When the value is 1, it means that the time management ability is optimal and the impact is the minimum value 1.

[0101] Taking into account the empowerment weight, the difficulty of knowledge point j and the mastery status value, the comprehensive priority of the knowledge point is calculated and expressed as:

[0102] ;

[0103] in Represented as the comprehensive priority of knowledge point j, It is expressed as the difficulty value of knowledge point j, represents the set of all previous knowledge points of knowledge point j, Indicates the mastery status value of learner i on the previous knowledge point u;

[0104] Taking into account the priority of knowledge point j and the total planned learning time, the target learning time of the knowledge point is allocated, which is expressed as:

[0105] ;

[0106] in represents the total learning time, M represents the total number of knowledge points, Represents the allocated learning time for knowledge point j.

[0107] By comprehensively considering the learner's mastery status value of the knowledge point, the learning time management value and the difficulty value of the knowledge point, the priority and target learning time of each knowledge point are calculated. The learning path can be dynamically adjusted according to the actual situation of each learner to achieve truly personalized learning path planning. The repeated learning weight of the knowledge point is adjusted by the learner's learning time management value of the knowledge point, which can effectively balance the learner's time allocation to each knowledge point and achieve self-improvement empowerment. In the target learning time allocation formula, the comprehensive priority and total learning time T of each knowledge point are combined to achieve scientific learning time allocation for each knowledge point. The higher the priority of the knowledge point, the more learning time will be allocated, while the time allocated to the knowledge point with lower priority will be reduced accordingly. By comprehensively calculating the learning priority and target learning time based on multiple factors such as mastery status, learning time management value and knowledge point difficulty, it can effectively prevent learners from being overly biased towards certain knowledge points and achieve the effect of ensuring the balance of the learning process.

[0108] The calculation formula for empowerment weights combines mastery status with time management. The weight calculation not only reflects the learner's mastery of knowledge points, but also comprehensively considers the learner's time management ability. A reasonable repeated learning weight is assigned to each knowledge point, which improves the personalization and effectiveness of the learning path.

[0109] The calculation formula for comprehensive priority is: The part represents the sum of the learner i's insufficient mastery of all the prerequisite knowledge points k for knowledge point j. Combined with the weight of the numerator and the multiplication of the difficulty of the knowledge point, the formula can dynamically adjust the priority of the current knowledge point. When the learner has a good grasp of the prerequisite knowledge points, the denominator is smaller, the priority is increased, and the learner can give priority to learning new knowledge points; otherwise, the priority is reduced, prompting the learner to review the prerequisite knowledge points first to ensure the coherence and completeness of the knowledge system.

[0110] The calculation formula for allocating study time takes into account the comprehensive priority and total study time to ensure that knowledge points with high priority are allocated more study time. More study time is allocated to high-priority knowledge points, which can ensure that learners have enough time to invest in learning important knowledge points. Less study time is allocated to low-priority knowledge points, which can avoid learners spending too much time on knowledge points that they have mastered well or have lower difficulty, thereby preventing waste of resources.

[0111] S3, define the objective function, use the multi-objective optimization algorithm to plan the learning path, construct the fusion population through non-dominated sorting, and iteratively select the optimal solution of the learning path.

[0112] Preferably, an objective function is defined, and a learning path is planned using a multi-objective optimization algorithm, including using a multi-objective optimization algorithm NSGA-II to dynamically plan the learner's learning path, initializing a population C and a population individual h, where the individual h represents a learning path and a corresponding learning time allocation scheme, and defining a learning path sequence and a learning time classification sequence for the population individual according to the comprehensive priority of the knowledge points and the allocation time, respectively, as expressed as:

[0113] ;

[0114] ;

[0115] in represents a sequence of learning paths, represents the learning time allocation sequence, represents the Vth knowledge point, Represents the Vth time allocation value.

[0116] Based on the mastery status value of knowledge point j by learner i in learning cycle t, the allocated learning time of knowledge point j and the comprehensive priority of knowledge point j, a specific time period in multiple learning is defined as learning cycle t. A learning cycle can be the sum of one day, one week, one learning module or one learning stage. The objective functions of maximizing learning effect and minimizing time allocation rationality are defined respectively, which are expressed as:

[0117] ;

[0118] ;

[0119] in represents the objective function of maximizing learning effect, represents the objective function of minimizing the rationality of time allocation, Represents the total number of knowledge points, and They represent the mastery status of learner i on knowledge point j in learning cycle t and learning cycle t-1 respectively, represents the learning priority of knowledge point k, Represents the learning time allocated to knowledge point j in learning cycle t.

[0120] Define the dominance relationship of individuals in a population, including the individuals in the population Less than or equal to individual , then it means that the individual Dominate , count the number of times individuals in the population are dominated; if individual x has no dominated individuals, then the individual is defined as non-dominated, and the set of all non-dominated individuals is ; For each pair of individuals in the population, the dominance relationship is determined, and the individuals in the population are constructed into the initial dominated set. The individuals with a dominated number of 0 are recorded as the first non-dominated set, and the individuals are removed from the population. The number of dominated times of the individuals in the dominated set is reduced by 1. If the number of dominated times of the individuals in the dominated set is reduced to 0, they are placed in the non-dominated set of the next layer. The process of placing individuals in different non-dominated sets is repeated to obtain a non-dominated ordering of the individuals.

[0121] Through the NSGA-II multi-objective optimization algorithm, the learner's learning path is dynamically planned, and the maximization of learning effect and the minimization of time allocation rationality are used as the optimization objective function, which realizes the comprehensive optimization of the learning path. According to the learner's mastery status, learning time allocation and knowledge point priority in different learning cycles, the learning path and learning time allocation plan are dynamically adjusted. While ensuring the learning effect, the learner's time allocation is optimized and the overall learning efficiency is improved. By taking the learner's learning performance in multiple learning cycles (such as mastery status value and learning time allocation) as input, it can provide customized learning path planning plans according to the learning needs of different learners and improve personalized learning effects.

[0122] By defining the dominance relationship of individuals in the population, the optimization of the learning path is divided into multiple non-dominated levels, and different individuals are divided into levels through non-dominated sorting. Non-dominated sorting can take into account both learning effect and the rationality of time allocation. Without sacrificing learning effect, it optimizes learning time allocation as much as possible, balances the conflicts between multiple objective functions, and provides learners with an optimal learning path with multi-objective balance. By defining the learning performance within the learning cycle t as the optimization variable, it dynamically adjusts the learners' mastery of knowledge points, learning time allocation plan and knowledge point priority according to different periods.

[0123] In the optimization objective function, not only the maximization of learning effect (i.e., the improvement of learners' mastery status in each learning cycle) is taken into consideration, but also the rationality of time allocation (i.e., the matching degree between actual learning time and priority) is taken into consideration. Through this multi-objective optimization, learners' time resources can be utilized to the maximum extent, avoiding over-learning and waste of learning resources, ensuring that learners obtain the best learning effect within a limited time. Through multi-objective optimization learning path planning, intelligent learning strategy adjustment plans can be provided according to learners' different learning goals (such as improving the mastery of a certain knowledge point or completing the learning of multiple knowledge points within a limited time).

[0124] Furthermore, a fusion population is constructed through non-dominated sorting, and the optimal solution of the learning path is iteratively selected, including taking individuals with low non-dominated layers based on non-dominated sorting as parent individuals, performing learning path crossover and learning time crossover on their learning path sequences and time allocation sequences to generate child individuals, wherein the learning path crossover is based on a randomly selected intersection point, and the learning path sequences of two parent individuals are exchanged before and after the intersection point to generate two new individuals, and the learning time crossover uses a linear weighted crossover method to mix the learning time allocation to generate a new learning time allocation.

[0125] Construct parent individuals and offspring individuals into a fusion population , and repeat the non-dominated sorting.

[0126] The current fusion population The individuals in the first non-dominated layer in are taken as the optimal solution set of the Pareto solution set, and the learning effect objective function is selected The solution with the maximum value is taken as the optimal solution of the learning path.

[0127] Apply the selected optimal solution for the learning path to the learner's next learning cycle.

[0128] By generating offspring individuals through learning path crossing and learning time crossing, the diversity of understanding is increased. Learning path crossing randomly selects intersection points and partially exchanges the learning paths of parent individuals to generate a new learning path sequence, so that the learner's learning path can cover more learning content and sequence. Learning time crossing uses linear weighted crossover method to integrate the learning time allocation of different parent individuals to generate a new time allocation plan, which can effectively prevent the algorithm from falling into local optimality and provide more solution space for learning path planning.

[0129] By performing multiple non-dominated sorting on the fusion population, the transmission of excellent genes can be guaranteed. At the same time, new solutions are continuously introduced to expand the diversity and scope of solutions, thereby improving the algorithm performance and accelerating the convergence to the Pareto frontier. The global optimality of multi-objective optimization is ensured by the Pareto optimal solution set. While ensuring the overall learning effect, the rationality of learning time allocation can be taken into account to achieve the global optimum. By applying the optimal solution of the learning path after each iteration to the learner's next learning cycle, the dynamic adjustment of the learning path can be realized, which can help learners continuously optimize learning strategies and improve learning effects. In the optimization process, maximizing the learning effect and the rationality of learning time allocation are used as two objective functions, which can effectively avoid the deviation caused by single-objective optimization. By comprehensively considering the balance between learning effect and time allocation, the priority and time allocation of the learning path can be flexibly adjusted.

[0130] S4, regularly calculates learners’ learning indicators, collects learners’ learning behavior data and mastery status to monitor learning progress.

[0131] Preferably, the learning indicators of the learners are calculated regularly, including testing the learning of the learners and calculating the learning indicators of the learners regularly, which can be expressed as:

[0132] ;

[0133] in represents the learning index of learner i, M represents the total number of knowledge points, represents the self-study test score of learner i on knowledge point j, Represents the highest score of knowledge point j.

[0134] The average learning index of all learners is used as the learning threshold. If the learner's learning index is lower than the learning threshold, it is judged to be lower than the learning expectation.

[0135] By regularly calculating learners' learning indicators and comparing them with the average learning indicators of all learners (learning threshold), automatic detection of learning status can be achieved, which can help learners understand their own learning progress and status in a timely manner, provide learners with timely reminders and guidance, and avoid a decline in learning effects. By regularly calculating learners' learning indicators and monitoring learners' learning status in real time, the learning system can realize intelligent monitoring and management of learners' learning process.

[0136] Further, collecting the learner's learning behavior data and mastery status to monitor the learning progress, including, based on the determination that the learning is below the learning expectations, further collecting the learner's learning behavior data and mastery status to monitor the learning progress, including the allocation and completion of learning time, the mastery status of knowledge points and the self-study test scores;

[0137] The time allocation and completion status, knowledge point mastery status and self-study test scores of all learners are counted separately, and based on the average of time allocation and completion status, knowledge point mastery status and self-study test scores, further self-learning evaluation is conducted on learners whose learning performance is below expectations. If any two of the time allocation and completion status, knowledge point mastery status and self-study test scores are below the average, the self-learning level is evaluated to be poor, and the path planning for the next learning cycle is re-performed.

[0138] By further collecting the learning behavior data and mastery status (such as learning time allocation, knowledge point mastery status and self-study test scores) of learners whose learning performance is below expectations, we can have a more comprehensive understanding of the specific problems of learners in the learning process. By comprehensively evaluating learners through three indicators, namely time allocation and completion status, knowledge point mastery status and self-study test scores, we can more accurately reflect the learners' learning situation. For learners who are assessed to have poor self-study, the system can promptly initiate a new round of learning path planning. For learners with poor learning levels, the system can provide them with personalized learning support plans in the next learning cycle, such as adjusting learning time allocation, increasing learning resources, arranging review plans, etc., which can not only improve learners' learning effects, but also improve the intelligence and personalization level of the learning system, and provide learners with more scientific and effective learning path planning and learning strategy support.

[0139] On the other hand, this embodiment also provides a self-learning and self-improvement weighted evaluation system driven by progressive planning effect data, which includes:

[0140] The stratification module obtains learners' self-study results and process parameters, analyzes the difficulty of knowledge points, and performs difficulty stratification based on the prerequisite relationship and difficulty change rate between knowledge points.

[0141] The statistical module counts learners’ self-study behavior to analyze the mastery status of knowledge points, assigns weights to knowledge points based on learning time and mastery status, analyzes the comprehensive priority of knowledge points and allocates learning time.

[0142] The optimization module defines the objective function, uses a multi-objective optimization algorithm to plan the learning path, constructs a fusion population through non-dominated sorting, and iteratively selects the optimal solution of the learning path.

[0143] The monitoring module regularly calculates learners’ learning indicators, collects learners’ learning behavior data and mastery status to monitor learning progress.

[0144] If the above 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 invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0145] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0146] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0147] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0148] Example 2 is an embodiment of the present invention, which provides a self-learning and self-improvement weighted evaluation method driven by progressive planning effect data. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0149] Three learners (Learners A, B, and C) were selected to conduct a four-week experiment on the course "Data Structure and Algorithm". Each learner will plan a personalized learning path based on the invented self-learning method in different learning cycles. The experiment will be divided into three learning cycles (T1, T2, and T3), and the learning content of each cycle will be dynamically adjusted according to the learner's learning progress and mastery status.

[0150] During the experimental preparation phase, the learning behavior data of each learner in different cycles are collected, including learning time, completion of learning tasks, test scores, video viewing times, number of discussions, etc. Based on these data, the system generates a personalized learning path and time allocation for each learner. The learning path optimization will be dynamically adjusted using a multi-objective optimization algorithm (such as NSGA-II) based on each learner's mastery of the knowledge and the difficulty of the knowledge points.

[0151] During the implementation process, learners’ learning paths in each time period (T1, T2, T3) are monitored and adjusted in real time to ensure that their learning effects and time allocation are optimized.

[0152] The control group used the traditional unified learning path design method, which does not dynamically adjust according to the learners' personalized data, but a fixed learning path and time allocation that applies to all learners. This method does not take into account the learners' mastery status and learning behavior, and the learning path is not optimized according to the progress of different learners.

[0153] Table 1 Experimental data table

[0154] It can be seen from the experimental data table that as each learner's learning cycle progresses, the learning time, test scores, video viewing times, discussion times, learning behavior values, and mastery status values ​​all show a gradual upward trend. This change reflects the actual effect of the method proposed in this invention in optimizing the learning path.

[0155] Specifically, the change in learning time reflects the time learners invest in each cycle. Learners A, B, and C spent 3.5 hours, 3.0 hours, and 2.8 hours learning in cycle T1, respectively, but in cycle T3, their learning time increased to 4.5 hours, 4.8 hours, and 4.0 hours, respectively. This shows that after systematic learning path optimization, learners can appropriately increase their learning time to improve their mastery level.

[0156] The average test scores also improved as the cycle progressed, with learner A improving from 80% in T1 to 90% in T3, learner B from 75% to 85%, and learner C from 70% to 80%. This improvement in performance shows that the invented learning path planning can effectively improve learners' mastery level, especially in the later stage, through the optimized allocation of learning time, learners can master knowledge more effectively.

[0157] The increase in video views and discussions reflects the depth and breadth of learners' input in the learning process. Learner A's video views increased from 10 to 15, and the number of discussions increased from 2 to 4 during the T1 to T3 period; learner B's video views increased from 8 to 14, and the number of discussions increased from 1 to 3; learner C's video views increased from 5 to 12, and the number of discussions increased from 3 to 5. These data show that with the optimization of the learning path, learners have become more proactive in the learning process, actively participated in video learning and discussions, and enhanced their understanding and application of knowledge.

[0158] The gradual increase in learning behavior value and mastery status value reflects the learning effect and knowledge mastery of learners in each cycle. For example, learner A's learning behavior value increased from 0.75 in T1 to 0.85 in T3, and the mastery status value increased from 0.78 to 0.88; learner B's learning behavior value increased from 0.70 to 0.80, and the mastery status value increased from 0.74 to 0.83; learner C's learning behavior value increased from 0.65 to 0.78, and the mastery status value increased from 0.68 to 0.80. This shows that with the continuous optimization of learning paths and time allocation, learners' learning status has been significantly improved.

[0159] Table 2 Data table of control group

[0160] The experimental group (the present invention) shows that the learning time of learners A, B, and C in each cycle gradually increases, and the test scores continue to improve with the adjustment of learning time and optimization path. For example, learner A improved from 80% in T1 to 90% in T3, while learner B improved from 75% to 85%. This phenomenon shows that the present invention promotes learners' in-depth learning in each learning cycle through personalized learning path optimization, thereby improving their scores.

[0161] In the control group (traditional method), all learners kept their study time the same and their test scores improved only slightly. For example, learner A's score only improved from 78% in T1 to 83% in T3, and learner B's score improved from 72% to 78%. This result shows that the traditional method lacks dynamic adjustment and personalized path planning, resulting in learners' progress being limited by a fixed learning path.

[0162] Learners in the experimental group, especially at T3, had a significant increase in video views and discussions. For example, learner A's video views increased from 10 to 15, and the number of discussions increased from 2 to 4. Learner B's video views increased from 10 to 14, and the number of discussions increased from 2 to 3. This shows that with the optimization of the learning path, learners not only invested more time in learning content, but also participated more actively in discussions during the learning process, which promoted the understanding and application of knowledge.

[0163] In the control group, the increase in video views and discussions was smaller, with learner A’s views increasing from 9 to 11 and discussions from 2 to 3, and learner B’s views increasing from 8 to 9 and discussions from 1 to 2. Due to the fixed learning path of the traditional method, learners’ learning enthusiasm and participation are relatively low.

[0164] The learning behavior value and mastery state value of the experimental group increased significantly in each cycle. The learning behavior value of learner A increased from 0.75 to 0.85, and the mastery state value increased from 0.78 to 0.88; the learning behavior value of learner B increased from 0.70 to 0.80, and the mastery state value increased from 0.74 to 0.83. This reflects that the present invention helps learners better master knowledge by dynamically adjusting the learning path and time allocation.

[0165] In the control group, the growth of learners' learning behavior value and mastery state value was small. For example, learner A's learning behavior value increased from 0.70 to 0.75, and the mastery state value increased from 0.75 to 0.79; learner B's learning behavior value increased from 0.65 to 0.70, and the mastery state value increased from 0.70 to 0.74. This shows that the traditional method lacks optimization for learners' personalized needs, resulting in limited improvement in learning effects.

[0166] Through the above comparative analysis, the experimental group (the present invention) shows obvious advantages, especially in terms of personalized learning path optimization, learning time allocation, learner enthusiasm and learning effect improvement, which are superior to the control group (traditional method). The present invention tailors the learning path for each learner by comprehensively considering the learner's learning time, mastery status and learning behavior, thereby significantly improving learning efficiency and learning effect.

[0167] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A self-learning and self-improvement weighted evaluation method driven by progressive planning effect data, characterized in that: include: Obtain learners' self-study results and process parameters, analyze the difficulty of knowledge points, and stratify the difficulty according to the prerequisite relationship and difficulty change rate between knowledge points; Statistics on learners’ self-study behavior are used to analyze the mastery of knowledge points. Based on the learning time and mastery status, weights are assigned to knowledge points to analyze the comprehensive priority of knowledge points and allocate learning time. Define the objective function, use the multi-objective optimization algorithm to plan the learning path, build the fusion population through non-dominated sorting, and iteratively select the optimal solution of the learning path; Regularly calculate learners' learning indicators, collect learners' learning behavior data and mastery status to monitor learning progress.

2. The progressive planning effect data-driven self-learning and self-improvement weighted evaluation method according to claim 1, characterized in that: The self-study results include: conducting self-study tests on self-study students based on the self-study question bank of the jth knowledge point, and calculating the average score of the students on the jth knowledge point; The process parameters include building a knowledge point self-study question bank and a self-study video bank based on self-study knowledge points, building a self-study group chat based on social software, and recording the number of students' discussions; Based on the self-study video library of the jth knowledge point, count the average number of views of students; Based on the jth knowledge point, count the average number of times students discuss the jth knowledge point; The difficulty of the knowledge point is calculated by the average test score, average number of views, and average number of discussions, expressed as: ； in, represents the difficulty of knowledge point j, represents the average score of the self-study test for knowledge point j, represents the average number of views of the self-study video library of knowledge point j, represents the average number of discussions on knowledge point j, , and They represent the lowest average score among all knowledge points, the highest average number of views among all knowledge points, and the highest average number of comments among all knowledge points respectively; Through the logical association of textbook materials and knowledge points, determine the knowledge points that must be learned before learning all knowledge points, and clarify the prerequisite relationship between knowledge points; Based on the prerequisite relationship of knowledge points, knowledge point j and all the prerequisite knowledge points are constructed into knowledge units. , and calculate the knowledge unit The average difficulty of all knowledge points ; According to the average difficulty Sort all knowledge units by size and calculate the average difficulty of all knowledge units The difficulty change rate between , and calculate the mean and standard deviation of all difficulty change rates. If the average difficulty of the knowledge unit If the difficulty change rate between them is greater than or equal to the sum of the mean and double standard deviation of all difficulty change rates, then it is judged that there is a difficulty gap; according to According to the size of the knowledge units, the knowledge points are sorted and the difficulty levels are stratified according to the difficulty gaps.

3. The progressive planning effect data-driven self-learning and self-improvement weighted evaluation method according to claim 2, characterized in that: The statistical learner's self-study learning behavior analysis knowledge point mastery status, Including, calculating the video learning completion rate based on the video viewing time and total time of the self-study video library of learner i for knowledge point j; Count the total learning time of learner i in learning knowledge point j, including the viewing time of the self-study video library, the reading time and the note-taking time; Count the number of discussions that learner i has on learning knowledge point j; According to the statistical data of learner i, including the video learning completion rate, the total learning time, the maximum video learning completion rate, learning time and discussion times of all learners, normalization is performed; The normalized video learning completeness rate, total learning time and the sum of discussion times are used as the learner's learning behavior data, and normalized according to the maximum learning behavior value of all learners i to obtain the learning behavior value of learner i for knowledge point j; Count the learner's abnormal behavior of watching videos in the self-study video library when learning knowledge point j, and count the number of non-learning related discussions when the learner is learning knowledge point j. Sum and normalize the abnormal behavior and non-learning related discussions; Based on the actual learning time of learner i for learning knowledge point j divided by the planned learning time, the learning time management value is calculated and normalized; The learner's mastery of knowledge points is calculated based on the learning behavior value, abnormal learning behavior value, learning time management value and self-study test average score, expressed as: ； in, represents the mastery status value of learner i on knowledge point j, represents the self-study test score of learner i on knowledge point j, represents the learning behavior value of learner i in the learning process of knowledge point j, represents the abnormal learning behavior value of learner i learning knowledge point j, It represents the learning time management value of learner i for learning knowledge point j; Count the historical average mastery status of all learners on all knowledge points, and calculate the overall standard deviation of the mastery status of all learners; If the learner's mastery status value of a knowledge point is greater than or equal to the sum of the historical average mastery status and the overall standard deviation of the mastery status, then the learner is judged to have a good grasp of the knowledge point; if the learner's mastery status value of a knowledge point is less than the sum of the historical average mastery status and the overall standard deviation of the mastery status, then the learner is judged to have a poor grasp of the knowledge point.

4. The progressive planning effect data-driven self-learning and self-improvement weighted evaluation method according to claim 3, characterized in that: The method of assigning weights to knowledge points based on learning time and mastery status, analyzing the comprehensive priority of knowledge points and allocating learning time includes calculating the repeated learning weight of knowledge point j according to the average value of the learner's learning time management value for different knowledge points, expressed as: ； in, represents the weight of knowledge point j, represents the average learning time management value of learner i; Taking into account the empowerment weight, the difficulty of knowledge point j and the mastery status value, the comprehensive priority of the knowledge point is calculated and expressed as: ； in, Represented as the comprehensive priority of knowledge point j, It is expressed as the difficulty value of knowledge point j, represents the set of all previous knowledge points of knowledge point j, Indicates the mastery status value of learner i on the previous knowledge point u; Taking into account the priority of knowledge point j and the total planned learning time, the target learning time of the knowledge point is allocated, which is expressed as: ； in, represents the total learning time, M represents the total number of knowledge points, Represents the allocated learning time for knowledge point j.

5. The progressive planning effect data-driven self-learning and self-improvement weighted evaluation method according to claim 4, characterized in that: Defining the objective function and using the multi-objective optimization algorithm to plan the learning path includes using the multi-objective optimization algorithm NSGA-II to dynamically plan the learner's learning path, initializing the population C and the population individual h, where the individual h represents the learning path and the corresponding learning time allocation plan; For individuals in the population, according to the comprehensive priority of knowledge points and allocate time The learning path sequence and learning time classification sequence are defined respectively, which are expressed as: ; ； in, represents a sequence of learning paths, represents the learning time allocation sequence, represents the Vth knowledge point, represents the Vth time allocation value; According to the learning cycle t in multiple learning, the objective functions of maximizing the learning effect and minimizing the rationality of time allocation are defined respectively, which are expressed as: ； ； in represents the objective function of maximizing learning effect, represents the objective function of minimizing the rationality of time allocation, Represents the total number of knowledge points, and They represent the mastery status of learner i on knowledge point j in learning cycle t and learning cycle t-1 respectively, represents the learning priority of knowledge point k, represents the learning time allocated to knowledge point j in learning cycle t; Define the dominance relationship of individuals in a population, including the individuals in the population Less than or equal to individual , then it means that the individual Dominate , count the number of times individuals in the population are dominated; If individual x has no dominant individuals, then the individual is defined as non-dominated. Let the set of all non-dominated individuals be ; The dominance relationship is determined for each pair of individuals in the population: the individuals in the population are constructed into the initial dominating set, the individuals with a dominance count of 0 are recorded as the first non-dominated set, the individuals are removed from the population, and the dominance count of the individuals in the dominating set is reduced by 1; if the dominance count of an individual in the dominating set is reduced to 0, it is placed in the non-dominated set of the next layer, and the process of placing the individuals in different non-dominated sets is repeated to obtain the non-dominated set sorting of the individuals.

6. The progressive planning effect data-driven self-learning and self-improvement weighted evaluation method according to claim 5, characterized in that: The method of constructing a fusion population through non-dominated sorting and iteratively selecting the optimal solution of the learning path includes: taking individuals with low non-dominated layers in the non-dominated sorting as parent individuals, performing learning path intersection and learning time intersection on the learning path sequence and the time allocation sequence to generate child individuals; Construct parent individuals and offspring individuals into a fusion population , and repeat the non-dominated sorting; The current fusion population The individuals in the first non-dominated layer in are taken as the optimal solution set of the Pareto solution set, and the learning effect objective function is selected The solution with the maximum value is taken as the optimal solution of the learning path; Apply the selected optimal solution for the learning path to the learner's next learning cycle.

7. The progressive planning effect data-driven self-learning and self-improvement weighted evaluation method according to claim 6, characterized in that: The periodic calculation of the learner's learning index includes detecting the learner's learning progress and periodically calculating the learner's learning index, which is expressed as: ； in represents the learning index of learner i, M represents the total number of knowledge points, represents the self-study test score of learner i on knowledge point j, represents the highest score of knowledge point j; The average learning index of all learners is used as the learning threshold. If the learner's learning index is lower than the learning threshold, it is judged as lower than the learning expectation. The collecting of the learner's learning behavior data and mastery status to monitor the learning progress includes, based on the determination that the learning result is lower than the learning expectation, further collecting the learner's learning behavior data and mastery status to monitor the learning progress, including the allocation and completion of learning time, the mastery status of knowledge points and the self-study test scores; The time allocation and completion status, knowledge point mastery status and self-study test scores of all learners are counted separately, and based on the average of time allocation and completion status, knowledge point mastery status and self-study test scores, further self-learning evaluation is conducted on learners whose learning performance is below expectations. If any two of the time allocation and completion status, knowledge point mastery status and self-study test scores are below the average, the self-learning level is evaluated to be poor, and the path planning for the next learning cycle is re-performed.

8. A self-learning and self-improvement weighted evaluation system driven by progressive planning effect data using the method according to any one of claims 1 to 7, characterized in that: The stratification module obtains learners' self-study results and process parameters, analyzes the difficulty of knowledge points, and stratifies the difficulty according to the prerequisite relationship and difficulty change rate between knowledge points; The statistical module counts learners’ self-study behavior to analyze the mastery of knowledge points, assigns weights to knowledge points based on learning time and mastery status, analyzes the comprehensive priority of knowledge points and allocates learning time; The optimization module defines the objective function, uses a multi-objective optimization algorithm to plan the learning path, constructs a fusion population through non-dominated sorting, and iteratively selects the optimal solution for the learning path; The monitoring module regularly calculates learners’ learning indicators, collects learners’ learning behavior data and mastery status to monitor learning progress.

9. A computer device comprising: A memory and a processor; the memory stores a computer program, wherein the processor implements the steps of any method as claimed in claim 1 when executing the computer program.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images