An adaptive programming practical training intelligent pushing method for high school information technology courses

Abstract

The application provides a self-adaptive programming practical training intelligent pushing method for a high school information technology course, and comprises the following steps: practical training behavior data collection, which is used for collecting multi-dimensional original behavior data in real time in the process of user programming practical training. The self-adaptive programming practical training intelligent pushing method for the high school information technology course provided by the application collects multi-dimensional behavior data of the whole process of user programming practical training, utilizes a high school information technology course exclusive knowledge graph, combines a time sequence iteration computer mechanism of a knowledge tracking model, updates the mastering probability of each knowledge point of the user in real time, simultaneously, adaptively adjusts the difficulty level and question type structure of the practical training task based on the real-time learning situation of the user, realizes hierarchical promotion training and targeted reinforcement training, avoids the problems of insufficient training targeting and poor adaptability, and improves the teaching level and intelligent education quality of the high school information technology programming practical training.

Description

Technical Field

[0001] This application relates to the field of online education technology, and in particular to an adaptive programming training intelligent push method for high school information technology courses. Background Technology

[0002] Programming practice in high school information technology courses is a core teaching component for cultivating users' computational thinking, logical reasoning, and programming skills. Compared with theoretical knowledge learning, programming practice is characterized by its strong practicality, complex logic, large individual differences, and diverse error types.

[0003] Current programming training models primarily involve reviewing basic syntax and knowledge points before class, followed by teachers demonstrating the entire process of writing, debugging, and running code step-by-step during class. Users then follow along to complete in-class exercises, culminating in a final homework assignment. However, this rote learning approach only allows users to briefly familiarize themselves with the steps, leaving knowledge points at the level of instantaneous memorization. It fails to build long-term knowledge reserves and programming thinking, and users easily forget these concepts after class. While users may appear to complete in-class exercises smoothly, they haven't truly mastered the knowledge and skills. This results in superficial knowledge acquisition, weak practical abilities, and an overall suboptimal programming training effect. Summary of the Invention

[0004] This application aims to at least partially solve one of the technical problems in the aforementioned technologies.

[0005] Therefore, one objective of this application is to propose an adaptive programming training intelligent push method for high school information technology courses. This method can adaptively adjust the difficulty level and question structure of training tasks based on the user's real-time learning situation, realize tiered and progressive training and targeted reinforcement training, avoid the problems of insufficient training relevance and poor adaptability, and improve the teaching level and intelligent education quality of high school information technology programming training.

[0006] To achieve the above objectives, the first aspect of this application proposes an adaptive programming training intelligent push method for high school information technology courses, comprising the following steps: training behavior data collection, used to collect multi-dimensional raw behavior data in real time during the user's programming training process, the multi-dimensional raw behavior data including login data, content interaction data, code editing behavior data, debugging and running behavior data, and evaluation practice result data; formulating a learning profile, based on the collected raw behavior data, using a knowledge graph of the high school information technology course as a framework to quantitatively analyze each knowledge point and obtain the user's real-time learning profile data; deep data analysis, based on the obtained learning profile data, using cognitive error classification mapping rules to perform error analysis and determine the weak knowledge points and cognitive roots corresponding to the errors; training task push, based on the judgment results, pushing gradient variation training tasks matching the current learning stage to the user; learning feedback, based on the completion results of the training and combined with the updated learning profile, formulating a learning feedback report, the feedback report including core cognitive problems, changes in knowledge point mastery, and subsequent learning suggestions.

[0007] In addition, the adaptive programming training intelligent push method for high school information technology courses proposed above in this application may also have the following additional technical features:

[0008] Specifically, the training behavior data collection steps include: acquiring code editing behavior data, debugging and running behavior data, and evaluation exercise result data when the user executes the operation instructions of the programming training task, and simultaneously recording login data and content interaction data; cleaning and preprocessing the collected multi-dimensional raw behavior data to remove invalid data.

[0009] Specifically, each programming training task is pre-set with multi-dimensional tags, which include corresponding knowledge point tags, difficulty level tags, curriculum standard academic quality level tags, and question type tags. The knowledge point tags are mapped one-to-one with the knowledge point entities in the knowledge graph of the high school information technology curriculum.

[0010] Specifically, the step of creating a learning profile includes: associating and matching preprocessed behavioral data with the knowledge point tags corresponding to the current programming training task and the knowledge point entities in a pre-set high school information technology course knowledge graph to obtain feature indicators corresponding to each knowledge point, including answer accuracy, answer time, number of debugging attempts, and error type distribution; based on the feature indicators, performing time-series iterative calculations through a pre-set knowledge tracking model to obtain the mastery probability value of each knowledge point; and generating a multi-dimensional real-time learning profile based on the knowledge point mastery probability and error feature data.

[0011] Specifically, the multi-dimensional real-time learning profile includes: a knowledge mastery layer, used to store the mastery probability, error distribution, and learning progress data of each knowledge point; a competency layer, used to record grammar proficiency, logical rigor, and program debugging ability; and a behavioral style layer, used to retain the proportion of thinking time, self-correction rate, and code writing rhythm feature tags.

[0012] Specifically, the deep data analysis step includes: classifying programming errors generated during practical training into different surface error types based on error feature data in the learning profile; mapping surface errors to corresponding cognitive roots through preset cognitive error classification mapping rules; determining the user's weak knowledge points and related knowledge points based on the prior association relationships of knowledge points in the knowledge graph of high school information technology courses, and outputting structured cognitive defect analysis results.

[0013] Specifically, the surface error types include syntax errors, logic errors, boundary condition errors, and algorithmic thinking errors, while the cognitive root causes include missing concept memories, misunderstanding of rules, logical thinking defects, and insufficient transfer and application abilities.

[0014] Specifically, the training task push step includes: based on the analysis results, matching the prerequisite knowledge points that are strongly related to the current knowledge point and generating corresponding question sets; based on the difficulty level label, question type label, and the knowledge point mastery probability of the user's learning profile, matching the gradient variation training tasks that are suitable for the user's current learning level from the training resource library, and arranging them in a progressive order to form a training sequence; comparing the training task completion results with the preset learning achievement threshold to determine whether the user's current knowledge point mastery level has reached the standard for entering the next stage of learning.

[0015] Specifically, determining whether a user's current level of mastery of a knowledge point meets the standard for entering the next stage of learning includes: when the completion result of the training task is greater than or equal to a preset learning achievement threshold, determining that the user's current level of mastery of the knowledge point meets the standard, allowing the unlocking of subsequent knowledge point training nodes and entry into the next stage of learning; when the completion result of the training task is less than the preset learning achievement threshold, determining that the user's current level of mastery of the knowledge point does not meet the standard, maintaining the current knowledge point training stage and continuing to strengthen learning.

[0016] Specifically, the learning feedback steps include: repeatedly executing the learning profile update and in-depth data analysis steps based on the results of the variation training tasks submitted by the user, comparing whether the errors in the variation training and the original errors belong to the same source cognitive bias; updating the mastery probability, error characteristics, and ability level data of the corresponding knowledge points in the learning profile in real time based on the results of the variation training tasks; generating a learning feedback report based on the updated learning profile; and simultaneously outputting overall class learning statistics to the teacher, including the class knowledge point mastery rate, the distribution of common error types, and the overall learning progress.

[0017] Compared with existing technologies, this application has the following beneficial effects: The adaptive programming training intelligent push method of this application for high school information technology courses collects multi-dimensional behavioral data of users throughout the entire programming training process, utilizes the exclusive knowledge graph of high school information technology courses, and combines the time-series iterative calculation mechanism of the knowledge tracking model to update the mastery probability of each knowledge point of the user in real time. It can comprehensively display the user's real learning status. At the same time, based on the user's real-time learning situation, it adaptively adjusts the difficulty level and question structure of the training tasks to achieve tiered advancement training and targeted reinforcement training, avoiding the problems of insufficient relevance and poor adaptability of training, and improving the teaching level and intelligent education quality of high school information technology programming training.

[0018] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0019] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0020] Figure 1 This is a schematic flowchart of an adaptive programming training intelligent push method for a high school information technology course according to an embodiment of this application. Detailed Implementation

[0021] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0022] The following describes an embodiment of the present application of an intelligent push method for adaptive programming training in a high school information technology course, with reference to the accompanying drawings.

[0023] like Figure 1As shown in the figure, an adaptive programming training intelligent push method for high school information technology courses according to an embodiment of this application includes the following steps:

[0024] Training behavior data collection is used to collect multi-dimensional raw behavior data in real time during users' programming training. The multi-dimensional raw behavior data includes login data, content interaction data, code editing behavior data, debugging and running behavior data, and evaluation exercise result data.

[0025] The specific steps for collecting training behavior data include: acquiring code editing behavior data, debugging and running behavior data, and evaluation exercise result data when users execute operation instructions for programming training tasks, and simultaneously recording login data and content interaction data; cleaning and preprocessing the collected multi-dimensional raw behavior data to remove invalid data.

[0026] Each programming training task is pre-set with multi-dimensional tags, including corresponding knowledge point tags, difficulty level tags, curriculum standard academic quality level tags, and question type tags. The knowledge point tags are mapped one-to-one with the knowledge point entities in the knowledge graph of the high school information technology course.

[0027] To illustrate with examples, firstly, login data includes basic information such as the user's account login time, login device, and training session duration. Content interaction data includes user browsing time of training questions, number of clicks on knowledge point materials, page switching operations, and training exit and re-entry records. Code editing behavior data records the entire process of user code input order, character deletion, statement modification, comment editing, and code formatting frame by frame. Debugging and running behavior data collects dynamic data such as program single run results, error types, number of active debugging attempts, retry interval after code modification, and error correction time. Evaluation practice result data summarizes the final answer correctness, score, incomplete test points, and record of incorrect questions in this training session.

[0028] After data collection is completed, valid behavioral data is retained by removing users' unconscious repeated enter key presses, meaningless garbled input, invalid operations from frequent undoing and redoing, empty data generated from long periods of inactivity, and redundant and duplicate data generated by lag. This ensures the accuracy and effectiveness of data for subsequent learning analysis. At the same time, through the pre-configuration of multi-dimensional tags, the collected user behavior data can be matched with knowledge points, difficulty, curriculum standard level, and question types, avoiding problems such as messy training data and inability to diagnose learning.

[0029] To create a learning profile, based on the collected raw behavioral data, quantitative analysis of each knowledge point is conducted using a knowledge graph of high school information technology courses as a framework to obtain real-time learning profile data of users.

[0030] The specific steps for creating a learning profile include: associating and matching preprocessed behavioral data with the knowledge point tags corresponding to the current programming training task and the knowledge point entities in the pre-set high school information technology curriculum knowledge graph to obtain feature indicators for each knowledge point, including answer accuracy, answer time, number of debugging attempts, and error type distribution; based on the feature indicators, performing time-series iterative calculations through a pre-set knowledge tracking model to obtain the mastery probability value for each knowledge point; and generating a multi-dimensional real-time learning profile based on the knowledge point mastery probability and error feature data.

[0031] The multi-dimensional real-time learning profile includes: a knowledge mastery layer, which stores the probability of mastering each knowledge point, error distribution, and learning progress data; a competency layer, which records grammar proficiency, logical rigor, and program debugging ability; and a behavioral style layer, which retains the percentage of thinking time, self-correction rate, and code writing rhythm characteristic tags.

[0032] To illustrate with examples, firstly, the accuracy rate is used to characterize the user's overall mastery of the knowledge points; the answering time reflects the user's familiarity with the knowledge points and the fluency of problem-solving; the number of debugging attempts reflects the user's code's fault tolerance and logical loopholes; and the error type distribution is used to identify the user's high-frequency error points and weak points. Furthermore, when using multi-dimensional feature indicators as input parameters and importing them into a pre-set knowledge tracking model for time-series iterative calculation, the model dynamically updates the data by combining the user's historical behavior data from previous training sessions with the current real-time training data to calculate the real-time mastery probability value of the knowledge points. The higher the probability value, the more mature the user's understanding and application ability of the corresponding knowledge points.

[0033] To illustrate, when the knowledge tracking model performs time-series iterative calculations, it retrieves the user's initial prior probability of mastering the corresponding knowledge points from their past experience. Then, it integrates in real time multi-dimensional feature indicators such as the correct answer rate, average answer time, number of debugging and corrections, error type distribution, and duration of thinking difficulties in the current training session. Through a built-in time-series weighting algorithm, it assigns higher influence weights to recent training behavior data and performs reasonable attenuation corrections on long-term historical learning data. It iteratively updates the posterior mastery probability of each knowledge point, and through multiple rounds of time-series iterations and repeated corrections, it finally outputs a dynamically updated knowledge point mastery probability value that fits the user's real-time learning status. This avoids the randomness of evaluation brought about by single training results and achieves continuous and dynamic quantitative evaluation of the user's knowledge point mastery.

[0034] In-depth data analysis, based on the acquired learning profile data, uses cognitive error classification mapping rules to analyze errors and identify the weak knowledge points and cognitive roots corresponding to the errors.

[0035] The in-depth data analysis steps specifically include: classifying programming errors generated during practical training into different surface error types based on error characteristic data in the learning profile; mapping surface errors to corresponding cognitive roots through preset cognitive error classification mapping rules; determining the user's weak knowledge points and related knowledge points based on the prior association relationships of knowledge points in the knowledge graph of high school information technology courses, and outputting structured cognitive defect analysis results.

[0036] Surface-level error types include syntax errors, logic errors, boundary condition errors, and algorithmic thinking errors. Cognitive root causes include missing concept memories, misunderstanding of rules, logical thinking defects, and insufficient ability to transfer and apply knowledge.

[0037] Using examples, all error records, debugging and modification paths, and error results generated by the user in this programming training are sorted out. According to the preset error judgment criteria, the problems that occurred during the user's programming process are divided into different surface error types. Among them, syntax errors refer to basic errors such as misspellings, incorrect use of keywords, non-standard punctuation, and missing statement formats. Logic errors refer to implicit errors such as disordered execution logic and execution results that do not match the expected results. Boundary condition errors refer to program loopholes such as the user's failure to consider extreme values, null values, critical intervals, and loop start and end conditions. Algorithm approach errors refer to structural program errors caused by deviations in the user's overall problem-solving framework and inappropriate selection of algorithm logic.

[0038] Based on this, the system calls the preset cognitive error classification mapping rules, which correspond to grammatical errors as missing concept memory, boundary condition errors as misunderstanding of rules, execution order errors as logical thinking defects, and algorithm architecture errors as insufficient transfer and application capabilities.

[0039] Training task push: Based on the judgment results, push graded variation training tasks that match the current learning stage to the user.

[0040] The specific steps for pushing practical training tasks include: Based on the analysis results, matching the prerequisite knowledge points that are strongly related to the current knowledge point and generating corresponding question sets; based on the difficulty level tags, question type tags, and the user's knowledge point mastery probability of the practical training task, matching the gradient variation practical training tasks that are suitable for the user's current learning level from the practical training resource library, and arranging them in a progressive order to form a practical training sequence; comparing the completion results of the practical training tasks with the preset learning achievement threshold to determine whether the user's current knowledge point mastery level has reached the standard for entering the next stage of learning.

[0041] The system determines whether a user's current level of mastery of a knowledge point meets the standard for entering the next stage of learning. This includes: when the completion result of the training task is greater than or equal to the preset learning achievement threshold, the user's current level of mastery of the knowledge point is determined to be up to standard, allowing the user to unlock subsequent knowledge point training nodes and enter the next stage of learning; when the completion result of the training task is less than the preset learning achievement threshold, the user's current level of mastery of the knowledge point is determined to be down to standard, maintaining the current knowledge point training stage and continuing to strengthen learning.

[0042] To illustrate with an example, firstly, by combining the difficulty level tags, question type tags, and curriculum standard academic quality level tags of the corresponding training tasks, the system selects graded variation training tasks from the pre-set training resource library that are suitable for the user's current learning level. It first pushes basic consolidation variation questions to correct the user's basic logical loopholes, and then pushes comprehensive application variation questions to train the user's ability to transfer knowledge across scenarios. The system is automatically sorted according to the progressive order of increasing training difficulty to avoid the inefficiency of training caused by uniform practice.

[0043] Learning feedback: Based on the completion results of the practical training and the updated learning profile, a learning feedback report is prepared. The feedback report includes core cognitive issues, changes in the mastery of knowledge points, and suggestions for subsequent learning.

[0044] The learning feedback process specifically includes: based on the results of the user-submitted variation training tasks, repeatedly executing the learning profile update and in-depth data analysis steps to compare whether the errors in the variation training and the original errors belong to the same source cognitive bias; based on the results of the variation training tasks, updating the mastery probability, error characteristics, and ability level data of the corresponding knowledge points in the learning profile in real time; generating a learning feedback report based on the updated learning profile; and simultaneously outputting overall class learning statistics to the teacher, including the class's knowledge point mastery rate, distribution of common error types, and overall learning progress.

[0045] To illustrate with an example, after a user completes a graded variation training task and submits their answer, the data, including the answer, is used as new input to repeatedly execute the dynamic update process of the learning profile and the in-depth cognitive deficiency data analysis process. By comparing the error characteristics that appeared in this training with the initial diagnostic error characteristics, it is determined whether the user still has homologous cognitive biases. If the number of similar boundary errors and logical nesting errors has decreased significantly, it is determined that the original cognitive shortcomings have been corrected. If the user still repeatedly makes homologous type errors, it is determined that the cognitive root cause has not been completely eradicated and is a weak point. By updating the corresponding dimension data of the learning profile in real time, the mastery probability of the corresponding knowledge points is dynamically adjusted up or down, and the error type distribution ledger is updated to ensure that the learning profile always remains consistent with the user's latest learning status.

[0046] In summary, this application presents an adaptive programming training intelligent push method for high school information technology courses. By collecting multi-dimensional behavioral data of users throughout the entire programming training process, utilizing a knowledge graph specific to high school information technology courses, and combining iterative calculation mechanisms of knowledge tracking models, the method updates the probability of users mastering each knowledge point in real time. This comprehensively displays the user's true learning status. Simultaneously, based on the user's real-time learning situation, the method adaptively adjusts the difficulty level and question structure of training tasks, achieving tiered advancement training and targeted reinforcement training. This avoids problems such as insufficient relevance and poor adaptability of training, thereby improving the teaching level and intelligent education quality of high school information technology programming training.

[0047] In the description of this specification, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0048] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0049] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for intelligent push notification of adaptive programming training in high school information technology courses, characterized in that, Includes the following steps: Training behavior data collection is used to collect multi-dimensional raw behavior data in real time during the user's programming training. The multi-dimensional raw behavior data includes login data, content interaction data, code editing behavior data, debugging and running behavior data, and evaluation exercise result data. To create a learning profile, based on the collected raw behavioral data, quantitative analysis of each knowledge point is conducted using a knowledge graph of high school information technology courses as a framework to obtain real-time learning profile data of users. In-depth data analysis, based on the acquired learning profile data, uses cognitive error classification mapping rules to analyze errors and determine the weak knowledge points and cognitive roots corresponding to the errors; Training task push: Based on the judgment result, push graded variation training tasks that match the current learning stage to the user; Learning feedback: Based on the completion results of the practical training and the updated learning profile, a learning feedback report is prepared. The feedback report includes core cognitive issues, changes in the mastery of knowledge points, and suggestions for subsequent learning.

2. The adaptive programming training intelligent push method for high school information technology courses according to claim 1, characterized in that, The specific steps for collecting training behavior data include: When users execute programming training tasks, the system acquires code editing behavior data, debugging and running behavior data, and evaluation exercise result data, and simultaneously records login data and content interaction data. The collected multi-dimensional raw behavioral data is cleaned and preprocessed to remove invalid data.

3. The adaptive programming training intelligent push method for high school information technology courses according to claim 2, characterized in that, Each programming training task is pre-set with multi-dimensional tags, which include corresponding knowledge point tags, difficulty level tags, curriculum standard academic quality level tags, and question type tags. The knowledge point tags are mapped one-to-one with the knowledge point entities in the knowledge graph of the high school information technology curriculum.

4. The adaptive programming training intelligent push method for high school information technology courses according to claim 3, characterized in that, The steps for creating a learning profile specifically include: The preprocessed behavioral data is associated and matched with the knowledge point tags corresponding to the current programming training task and the knowledge point entities in the pre-set high school information technology course knowledge graph to obtain the feature indicators corresponding to each knowledge point. The feature indicators include the answer accuracy rate, answer time, number of debugging times and error type distribution. Based on the aforementioned feature indicators, a time-series iterative calculation is performed using a pre-set knowledge tracking model to obtain the mastery probability value of each knowledge point. Based on the probability of mastering knowledge points and error characteristics data, a multi-dimensional real-time learning profile is generated.

5. The adaptive programming training intelligent push method for high school information technology courses according to claim 4, characterized in that, The multi-dimensional real-time learning profile includes: The knowledge mastery layer is used to store data on the mastery probability, error distribution, and learning progress of each knowledge point. The competency and skills layer is used to record grammatical proficiency, logical rigor, and program debugging ability. The behavioral style layer is used to retain the percentage of time spent thinking, the rate of self-correction, and the characteristic tags of coding rhythm.

6. The adaptive programming training intelligent push method for high school information technology courses according to claim 5, characterized in that, The deep data analysis steps specifically include: Based on the error characteristic data in the learning profile, the programming errors generated during the practical training process are classified into different surface error types; By using a pre-defined cognitive error classification mapping rule, surface errors are mapped to their corresponding cognitive root causes. Based on the prior relationships of knowledge points in the knowledge graph of high school information technology courses, the user's weak knowledge points and related knowledge points are identified, and structured cognitive deficit analysis results are output.

7. The adaptive programming training intelligent push method for high school information technology courses according to claim 6, characterized in that, The surface error types include syntax errors, logic errors, boundary condition errors, and algorithmic thinking errors. The cognitive root causes include missing concept memories, misunderstanding of rules, defects in logical thinking, and insufficient ability to transfer and apply knowledge.

8. The adaptive programming training intelligent push method for high school information technology courses according to claim 7, characterized in that, The specific steps for pushing out the practical training tasks include: Based on the analysis results, match the prerequisite knowledge points that are strongly related to the current knowledge point, and generate corresponding question sets; Based on the difficulty level tags, question type tags, and knowledge point mastery probability of the training tasks and the user's learning profile, the training resource library is used to match the gradient variation training tasks that are suitable for the user's current learning situation and arrange them in a progressive order to form a training sequence. The results of completing the practical training tasks are compared with the preset learning achievement thresholds to determine whether the user's current level of mastery of knowledge points has reached the standard for entering the next stage of learning.

9. The adaptive programming training intelligent push method for high school information technology courses according to claim 8, characterized in that, Whether a user's current level of mastery of a knowledge point meets the standard for entering the next stage of learning includes: When the completion result of the training task is greater than or equal to the preset learning achievement threshold, it is determined that the user's current knowledge point mastery level meets the standard, allowing the unlocking of subsequent knowledge point training nodes and entry into the next stage of learning. When the completion result of the practical training task is less than the preset learning achievement threshold, it is determined that the user's current knowledge level has not met the standard, the current knowledge training stage is maintained and learning is further strengthened.

10. The adaptive programming training intelligent push method for high school information technology courses according to claim 9, characterized in that, The learning feedback steps specifically include: Based on the results of the variation training tasks submitted by users, repeat the steps of updating the learning profile and conducting in-depth data analysis to compare whether the errors in the variation training and the original errors belong to the same source cognitive bias. Based on the results of the variation training tasks, the mastery probability, error characteristics and ability level data of the corresponding knowledge points in the learning profile are updated in real time. Based on the updated learning profile, a learning feedback report is generated; The system simultaneously outputs overall class learning statistics to teachers, including the class's mastery rate of knowledge points, the distribution of common error types, and the overall learning progress.

Images