Question pushing method and system based on intelligent learning system, medium and product

Abstract

The invention discloses a question pushing method and system based on an intelligent learning system, a medium and a product, and relates to the field of knowledge maps. The method comprises the following steps: constructing an initial exercise question sequence consisting of mother questions and deformation questions from a preset question bank containing the mother questions, the deformation questions and the migration questions; when the user answers the target question correctly, the question is removed from the sequence, and a migration question associated with the question is acquired and added to the sequence based on a preset data link relation, so that the exercise content is dynamically updated; when the user answers the target questions wrongly, the target questions are placed into the sequence again. According to the method, a dynamic question pushing method is constructed, a good cognitive state after the user answers the question is utilized, the user is actively guided to carry out knowledge expansion and application, the learning bottleneck caused by'stopping after answering 'in related technologies is effectively solved, single practice becomes deep learning experience integrating consolidation, expansion and application, and the learning experience is improved. And the knowledge migration capability of the user is improved.

Description

Technical Field

[0001] This application relates to the field of knowledge graphs, and in particular to a question-reasoning method, system, medium, and product based on an intelligent learning system. Background Technology

[0002] With the rapid development of information technology and the widespread adoption of the internet, online education has become an important component of the modern education system. In intelligent learning systems, question recommendation is a crucial step for users to consolidate knowledge and assess learning outcomes. A scientific and efficient question recommendation mechanism is essential for improving learners' ability to apply knowledge.

[0003] Currently, a common question recommendation method used in intelligent learning systems is as follows: When the system detects that a user has completed a course chapter or knowledge point, it retrieves practice questions matching the tags of that knowledge point from the question bank and pushes them to the user. The system records the user's answers, especially for questions answered incorrectly. When a user answers incorrectly, the system usually adds the wrong question or similar questions with the same knowledge point to the user's "Error Notebook" or "Needs Strengthening" list for later focused review or to be pushed again at a future time, thus achieving repeated reinforcement of weak areas. When a user answers a question correctly, the system considers that the user has initially mastered the knowledge point, and will reduce or stop pushing related questions, instead guiding the user to learn the next knowledge point.

[0004] However, the solutions offered by these technologies primarily focus on correcting and remedying incorrect answers, neglecting to deepen and expand upon correct responses. When a user answers a question correctly, indicating a basic grasp of the corresponding knowledge point, the system interrupts further learning on that topic. This approach fails to capitalize on the positive cognitive state demonstrated by a correct answer to guide the user towards further extended learning and enhance their understanding of the knowledge point, thus reducing the user's learning efficiency. Summary of the Invention

[0005] This application provides a question-reasoning method, system, medium, and product based on an intelligent learning system, which is used to improve users' ability to transfer knowledge points and their learning efficiency.

[0006] Firstly, this application provides a question-reasoning method based on an intelligent learning system, applied to the intelligent learning system, comprising: when a learning node of a course is displayed as completed, extracting the mother question corresponding to the knowledge point of the learning node from a preset question bank, wherein the preset question bank includes at least mother questions and variant questions that are interconnected through a first data link relationship, and transfer questions that are interconnected through a second data link relationship with the mother questions or variant questions; based on the first data link relationship, obtaining variant questions associated with the mother questions, and combining the mother questions and variant questions to form a sequence of practice questions, wherein the variant questions are questions that examine the same knowledge points as the mother questions but have different presentation formats; if the sequence of practice questions is not empty, then the practice questions are... The first question in the sequence is pushed to the user terminal as the target question, and the system receives the answer from the user terminal. If the answer is incorrect, the target question is added to the first preset position in the practice question sequence to update the practice question sequence. If the answer is correct, the target question is removed from the practice question sequence. Based on the second data link relationship, the system obtains the transfer questions associated with the target question. The transfer questions are added to the second preset position in the practice question sequence to update the practice question sequence. The transfer questions are questions that test knowledge points related to the target question. After updating the practice question sequence according to the answer, if the practice question sequence is empty, the practice ends.

[0007] By adopting the above technical solution, this approach constructs a dynamic feedback-based closed-loop practice system, overcoming the limitations of related technologies that stop once a correct answer is obtained. First, the intelligent learning system pre-assembles a sequence of practice questions, including core questions and variations, ensuring users' basic mastery and flexible understanding of core knowledge points. Furthermore, when a user answers correctly, the intelligent learning system does not terminate the practice but leverages the user's positive cognitive state by pushing transfer questions that test related knowledge points, guiding the user to extend their learned knowledge to broader application scenarios. Finally, the intelligent learning system dynamically updates the practice question sequence based on the answer results, making the practice process more adaptive and continuous. In summary, this solution transforms a single practice session from a simple knowledge point test into a deep learning experience integrating consolidation, expansion, and application, effectively improving users' knowledge transfer ability and comprehensive application level, thereby enhancing their learning efficiency.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, before the step of removing the target question from the practice question sequence, the method further includes: acquiring answering process data of the target question, the answering process data including answering time, answering step trajectory, and historical answering frequency; determining a first mastery score based on the ratio of the answering time to the preset standard time of the target question; determining a second mastery score based on the historical answering frequency and question type; determining a third mastery score based on the degree of matching between the answering step trajectory and the standard problem-solving steps of the target question; weighted summing of the first mastery score, the second mastery score, and the third mastery score to obtain a comprehensive mastery score; if the comprehensive mastery score is greater than or equal to the preset mastery score, then the step of removing the target question from the practice question sequence is performed; if the comprehensive mastery score is less than the preset mastery score, and the question type is not a transfer question, then a variant question associated with the target question is acquired, and the variant question is added to the head position of the practice question sequence.

[0009] By adopting the above technical solution, this approach solves the problem that a binary answer of correct or incorrect cannot accurately determine a user's true level of mastery. The intelligent learning system constructs a comprehensive mastery score by comprehensively analyzing answering time, answering steps, and historical answer counts. This allows the system to effectively distinguish between lucky correct answers and genuine mastery. Furthermore, when the comprehensive mastery score is below the required level, the system pushes relevant variation questions to the front of the queue for immediate and targeted practice, quickly addressing any gaps in the user's understanding. Therefore, this question-pushing method ensures that a question is only removed when the user is proven to have truly mastered it, avoiding learning gaps caused by false mastery and thus improving the reliability of knowledge consolidation.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, after the step of removing the target question from the practice question sequence, the method further includes: obtaining the base score corresponding to each question type from a preset question bank, and obtaining the user's answer accuracy rate for each question, wherein the base score of the transfer question is greater than the base score of the variant question, and the base score of the variant question is greater than the base score of the original question; calculating the user's practice score at the current learning node based on the base score and the answer accuracy rate; when the practice score is greater than or equal to the preset passing score, and the practice question sequence is not empty, sending a learning selection prompt to the user terminal, the learning selection prompt including continuing practice or entering the next learning node; when the selection information returned by the user terminal is to continue practice, keeping the state of the practice question sequence unchanged; when the selection information returned by the user terminal is to enter the next learning node, clearing the practice question sequence.

[0011] By adopting the above technical solution, the intelligent learning system first quantifies users' learning outcomes by setting progressively increasing base scores for different question types. This also positively incentivizes users to challenge themselves with more difficult questions, thereby pursuing deeper knowledge mastery. When a user's practice score reaches the preset passing score, the intelligent learning system does not forcibly end the practice but provides learning options, fully respecting the user's learning intentions and pace. Users with extra capacity can continue to deepen their understanding; those wishing to move to the next learning stage can smoothly transition. Therefore, this solution ensures users achieve their basic learning goals while providing personalized learning freedom, effectively enhancing user initiative and engagement, and optimizing the overall learning experience.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, based on the second data link relationship, acquiring transfer questions associated with the target question specifically includes: acquiring the answer process data, knowledge point data, and a set of candidate transfer questions that have a second data link relationship with the target question; extracting the question attribute data of each transfer question in the candidate transfer question set, wherein the question attribute data includes at least the difficulty coefficient, the type of ability being tested, and knowledge point data; determining the knowledge point matching degree based on the degree of overlap between the knowledge point data of the target question and the knowledge point data of each transfer question; and selecting the final transfer question from the candidate transfer question set based on the answer process data, question attribute data, and knowledge point matching degree.

[0013] By adopting the above technical solution, this approach achieves a personalized and precise selection strategy for transfer questions, ensuring the effectiveness and suitability of knowledge expansion. The intelligent learning system comprehensively utilizes user answer process data, question attribute data, and knowledge point matching to conduct a comprehensive evaluation. Furthermore, the intelligent learning system can accurately select the final transfer question from the candidate transfer question set that best matches the user's current cognitive level and effectively promotes their knowledge development. Therefore, this solution achieves the scientific nature and effectiveness of knowledge transfer decision-making in the question recommendation method, improving the success rate and efficiency of extended learning.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, adding a transfer question to a second preset position in the practice question sequence to update the practice question sequence specifically includes: obtaining the total number of questions in the practice question sequence and calculating the sequence length value; pausing the addition of transfer questions when the sequence length value is greater than or equal to a preset queue length threshold; obtaining the difficulty coefficient of each question in the practice question sequence and calculating the average difficulty value of the sequence when the sequence length value is less than the preset queue length threshold; obtaining the difficulty coefficient of the transfer question to be added; adding the transfer question to the end position of the practice question sequence when the difficulty coefficient of the transfer question is greater than the average difficulty value of the sequence; and adding the transfer question to the beginning position of the practice question sequence when the difficulty coefficient of the transfer question is less than or equal to the average difficulty value of the sequence.

[0015] By adopting the above technical solution, the intelligent learning system first checks the sequence length when deciding to add transfer questions to the practice question sequence. This avoids user anxiety caused by an excessively long to-do list, providing a fundamental guarantee for the user's learning experience. Furthermore, the intelligent learning system dynamically determines the insertion position of transfer questions based on a comparison between the difficulty coefficient of the transfer question and the average difficulty value of the sequence. Placing more difficult transfer questions at the end allows users to smoothly transition and gradually meet the challenge; while placing transfer questions of similar or lower difficulty at the beginning enables immediate consolidation and learning while the momentum is strong. Therefore, this solution makes the difficulty changes in the practice question sequence smoother and more logical, avoiding learning interruptions caused by sudden increases or decreases in difficulty, thereby improving the efficiency and smoothness of the practice process.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, after obtaining the difficulty coefficient of each question in the practice question sequence and calculating the average difficulty value of the sequence when the sequence length value is less than a preset queue length threshold, the method further includes: obtaining the estimated time for each question in the practice question sequence, and calculating the sum of the estimated time for each question in the practice question sequence to obtain the total time of the sequence, wherein the estimated time is the baseline time required to complete the questions marked in the preset question bank; obtaining the practice load value based on the product of the total time of the sequence and the average difficulty value of the sequence; pausing the addition of transfer questions when the practice load value is greater than or equal to a preset load threshold; and executing the step of obtaining the difficulty coefficient of the transfer questions to be added when the practice load value is less than the preset load threshold.

[0017] By adopting the above technical solution, this approach, based on queue length control, constructs a scientific practice load index by combining the total sequence time (time cost) and the average sequence difficulty (intellectual cost), realistically reflecting the overall cognitive pressure of the current practice task on the user. When the practice load exceeds the preset load threshold, the intelligent learning system will pause adding new transfer questions, adding an "overload protection" to the user's learning process. This proactively avoids learning fatigue and efficiency decline caused by continuous high-difficulty, high-time-consuming practice, thereby ensuring the long-term learning effect and user motivation.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, before adding the target question to the first preset position of the practice question sequence, the method further includes: acquiring incorrect answer process data of the target question, the incorrect answer process data including the user's answer steps and intermediate calculation results; acquiring standard answer process data of the target question in a preset question bank, the standard answer process data including a standardized sequence of problem-solving steps and the correct result corresponding to each step; determining answer error characteristics based on the incorrect answer process data and the standard answer process data, and obtaining the error cause based on the answer error characteristics; generating feedback information based on the error cause, and sending the feedback information to the user terminal, the feedback information including at least the error cause and correction suggestions.

[0019] By employing the aforementioned technical solution, the intelligent learning system acquires data on users' incorrect answer processes and compares it with standard answer process data. This allows for precise identification of the specific steps and nature of the errors, thus determining the characteristics of the errors and deriving the underlying causes, such as unclear concepts or calculation oversights. Based on these causes, the intelligent learning system automatically generates personalized feedback information containing corrective suggestions, providing immediate, diagnostically valuable feedback and a clear path for improvement. Therefore, this solution improves the efficiency and effectiveness of error correction, transforming each mistake into a valuable learning opportunity, thereby enhancing users' mastery of weak knowledge points.

[0020] Secondly, this application provides an intelligent learning system, which includes: one or more processors and a memory; the memory is coupled to one or more processors, and the memory is used to store computer program code, which includes computer instructions, and the one or more processors call the computer instructions to cause the intelligent learning system to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, this application provides a computer-readable storage medium including instructions that, when executed on an intelligent learning system, cause the intelligent learning system to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, this application provides a computer program product that, when run on an intelligent learning system, causes the intelligent learning system to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the intelligent learning system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. By adopting a technical solution that retrieves transfer questions related to the target question when the answer is correct and adds them to the practice question sequence, the system can take advantage of the user's good cognitive state after answering correctly to actively guide them to expand their knowledge and practice application. This effectively solves the problem in related technologies where practice ends as soon as a question is answered correctly, resulting in insufficient learning depth and failure to promote knowledge transfer. In this way, it transforms a single practice session from a knowledge point test into a deep learning experience that integrates consolidation, expansion, and application, thereby improving the user's knowledge transfer ability.

[0025] 2. By employing a technical solution that uses data from the answering process to calculate a comprehensive mastery score, and then comparing this score with a preset mastery score to determine whether to remove a target question or add a variation of the question, the system can accurately distinguish between a user's "true mastery" and "lucky correct answers." This effectively solves the problem in related technologies where relying solely on binary answer results can easily lead to learning gaps due to "false mastery." Consequently, the system achieves a refined assessment and targeted reinforcement of the user's knowledge mastery level, significantly improving the solidity of knowledge acquisition.

[0026] 3. By adopting a technical solution that checks the sequence length before adding transfer questions and determines whether to add them to the beginning or end of the queue by comparing the difficulty coefficient of the transfer questions with the average difficulty value of the sequence, the system can dynamically adjust the length and difficulty of the practice queue. This effectively solves the problem in related technologies where question recommendations may cause user anxiety due to an excessively long queue or disrupt the learning rhythm due to excessively large jumps in difficulty. As a result, the system achieves refined management of the user's learning load and cognitive rhythm, ensuring a smooth, efficient, and sustainable practice experience. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating a problem-solving method based on an intelligent learning system in an embodiment of this application; Figure 2This is another flowchart illustrating a problem-solving method based on an intelligent learning system in an embodiment of this application; Figure 3 This is a schematic diagram of the physical device structure of an intelligent learning system in the embodiments of this application. Detailed Implementation

[0028] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0029] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0030] The following describes the process of the method provided in this implementation. Please refer to [link / reference]. Figure 1 This is a flowchart illustrating a problem-solving method based on an intelligent learning system in an embodiment of this application.

[0031] S101. When the learning node of the course is displayed as complete, extract the mother question corresponding to the knowledge point of the learning node from the preset question bank. The preset question bank includes at least mother questions and variant questions that are interconnected through a first data link relationship, as well as transfer questions that are interconnected with mother questions or variant questions through a second data link relationship. In this context, a learning node represents the smallest structured learning unit in an online course, typically corresponding to a specific knowledge concept, a teaching video, or a chapter. It serves as the basic unit for the intelligent learning system to track a user's learning progress, such as "Chapter 1, Section 2: The Formula for Solving Quadratic Equations." A pre-set question bank is a structured database containing a large number of questions linked by predefined relationships, supporting practice and assessment functions. A master question is the most representative and fundamental question for a specific knowledge point, directly and clearly assessing the user's basic understanding and mastery of that knowledge point. A variation question, while maintaining the core knowledge point identical to the master question, is generated by changing the question's background, question format, data, or known conditions, assessing the depth and flexibility of the user's understanding of the knowledge point. A transfer question tests knowledge points related to the target question but may involve new knowledge points, combinations of multiple knowledge points, or applications in entirely new contexts. The first data link relationship connects a master question with all its variation questions in the pre-set question bank, forming a question cluster. The second data link relationship is used to connect a question (original question or variant question) with a transfer question that tests related knowledge points. The first and second data link relationships in the question bank can be realized through relationship edges such as "hasVariant" (having variants) and "leadsToTransfer" (leading transfer) in the knowledge graph.

[0032] When a user completes a learning node in the intelligent learning system, the system triggers a practice session to reinforce what they have learned. Specifically, the intelligent learning system first extracts the core knowledge point identifier corresponding to the learning node the user just completed. Then, using this knowledge point identifier as a search criterion, the system queries the pre-set question bank for parent question types and extracts parent questions directly related to that knowledge point.

[0033] S102. Based on the first data link relationship, obtain the variant questions associated with the mother question, and combine the mother question and the variant questions to form a practice question sequence. The variant questions are questions that test the same knowledge points as the mother question but are presented in a different form. The practice question sequence refers to a dynamically generated, ordered list of questions for the user. The intelligent learning system will push questions to the user in the order of this sequence. It is usually managed by a queue data structure, which has the characteristic of "first-in, first-out".

[0034] The intelligent learning system executes this step after successfully extracting the parent question from the preset question bank for the current learning node. Specifically, after obtaining the parent question, the intelligent learning system does not immediately push it to the user, but first constructs a practice sequence. The intelligent learning system uses the first data link relationship defined in the preset question bank to query all variant questions associated with the extracted parent question. This query process can be understood as starting from the parent question node in the knowledge graph and traversing all adjacent nodes pointed to by the first data link relationship. After obtaining all variant questions, the intelligent learning system integrates the parent question and these variant questions together to form an initial sequence of practice questions. Usually, to ensure the logical order of learning, the parent question is placed at the head of the practice question sequence, while the variant questions are arranged after the parent question according to preset rules (such as increasing difficulty, creation time, or randomness).

[0035] S103. If the sequence of practice questions is not empty, the first question in the sequence of practice questions will be pushed to the user terminal as the target question, and the answer returned by the user terminal will be received. The "non-empty" condition in the practice question sequence is a check to see if there are still questions in the current list of questions to be practiced for the user. This is crucial for controlling whether the practice loop continues. The "target question" is a temporary designation representing a question that has been pushed to the user and is awaiting their response; its content is taken from the first question in the practice question sequence. The "response result" refers to the data returned by the user's terminal to the intelligent learning system's backend after the user completes the question. The most basic result is a "correct" or "incorrect" judgment, but it can also include richer information such as the user's final answer, the time taken to complete the task, and detailed solution steps.

[0036] This step is executed by the intelligent learning system after constructing the sequence of practice questions and when the sequence is not empty. This step describes how the intelligent learning system pushes questions and obtains user feedback. Specifically, the intelligent learning system first checks the current state of the sequence of practice questions. If there is at least one question in the sequence (i.e., the sequence is not empty), the intelligent learning system reads the data of the first question in the queue and designates it as the "target question" for this interaction. Then, the intelligent learning system sends the content of the target question to the user's terminal and listens for the answer to the target question from the user's terminal. When the user submits an answer on the terminal, the terminal packages the answer and sends it back. The intelligent learning system receives and parses the result to prepare for subsequent logical judgments (S104 or S105).

[0037] S104. When the answer is incorrect, add the target question to the first preset position in the practice question sequence to update the practice question sequence; The first preset position refers to a specific position reserved in the sequence of practice questions for questions answered incorrectly.

[0038] The intelligent learning system executes this step when it receives the user's answer and determines that the answer is incorrect. Specifically, when the intelligent learning system detects an incorrect answer, it temporarily removes the incorrectly answered question from its current position in the practice question sequence. Then, according to a preset error handling strategy, the intelligent learning system re-inserts the question into the first preset position in the practice question sequence. The first preset position can be set using the following scheme: calculating the current length of the practice question sequence, and then determining the position one-quarter of the sequence length from the head of the queue as the first preset position. This position setting references the interval effect theory, aiming to promote the formation of long-term memory when the user encounters the question again while still remembering the error but not immediately redoing it. After re-inserting the question, the intelligent learning system recalculates and adjusts the state of the practice question sequence to prepare for the next round of push.

[0039] Understandably, in other embodiments, the first preset position may also be the end position of the practice question sequence to ensure that the user reviews after completing other questions; or it may be an optimal review position dynamically calculated based on the number of errors and importance of the question. This application aims to postpone redoing incorrect questions, rather than being limited to a specific position.

[0040] S105. When the answer is correct, remove the target question from the practice question sequence; based on the second data link relationship, obtain the transfer question associated with the target question; add the transfer question to the second preset position of the practice question sequence to update the practice question sequence. The transfer question is a question that examines the knowledge points associated with the target question. The second preset position refers to the position reserved for transfer questions in the sequence of practice questions. Its setting needs to take into account the user's cognitive load and learning pace.

[0041] The intelligent learning system executes this step upon receiving the user's answer from the user's terminal and determining that the answer is correct. Specifically, after confirming that the user has answered correctly, the intelligent learning system first permanently removes the target question from the practice question sequence, indicating that the user has mastered the knowledge point tested by the question. Next, the intelligent learning system retrieves transfer questions associated with the target question based on the second data link relationship in the preset question bank. After obtaining a suitable transfer question, the intelligent learning system adds it to the second preset position in the sequence according to the current state of the practice question sequence and the characteristics of the transfer question (such as difficulty, knowledge point coverage, etc.).

[0042] S106. After updating the practice question sequence based on the answer results, if the practice question sequence is empty, then end the practice.

[0043] "End practice" means that the intelligent learning system terminates the practice session at the current learning node, which may trigger subsequent processes such as performance statistics and learning report generation.

[0044] After updating the sequence of practice questions, the intelligent learning system performs this step to determine whether the practice should end. Specifically, after each update, the system immediately checks the current state of the sequence. If the sequence is empty, it means the user has either correctly completed all questions (including existing questions and dynamically added practice questions) or triggered a termination condition that cleared the sequence. In this case, the system initiates the practice termination process, which may include generating a practice report, calculating mastery metrics, and saving learning records. This check ensures that the practice session ends naturally at the appropriate time, avoiding infinite loops or exhaustion of practice content.

[0045] In this embodiment, by employing a technical solution that constructs an initial sequence of practice questions using master questions and variations, and dynamically updates the sequence by acquiring associated transfer questions when the user answers correctly, the user's learning process can be transformed from a passive test of a single knowledge point into an active deepening process that integrates basic consolidation, flexible understanding, and knowledge transfer. This effectively solves the problem in related technologies where the question-pushing mechanism only focuses on correcting wrong questions and ignores deepening and expanding correct answers, resulting in limited learning efficiency and knowledge application ability. Thus, a dynamic closed-loop, spiral-upward intelligent practice mode is achieved, enhancing the user's knowledge transfer ability and comprehensive quality.

[0046] Based on the above embodiments, the method provided in this embodiment will be described in further detail below. Please refer to... Figure 2 This is another flowchart illustrating a problem-solving method based on an intelligent learning system in an embodiment of this application.

[0047] S201. If the sequence of practice questions is not empty, push the first question of the sequence as the target question to the user terminal and receive the answer returned by the user terminal. Step S201 and Figure 1 The description of step S103 in the above embodiments is similar and will not be repeated here.

[0048] S202. When the answer result is incorrect, obtain the incorrect answer process data of the target question. The incorrect answer process data includes the user's answer steps and intermediate calculation results. The error response process data represents all the actions and results generated by the user during the process of answering the target question, used to analyze the user's problem-solving approach and the reasons for errors. The answer steps refer to each specific operation and calculation step in the problem-solving process, recorded chronologically to reflect the user's problem-solving path. Intermediate calculation results represent the various stage results generated by the user during the problem-solving process, including intermediate values ​​such as formula transformations and numerical calculations, used to pinpoint the specific location of calculation errors.

[0049] The intelligent learning system executes this step when it receives a user's answer and determines it to be incorrect. Specifically, the intelligent learning system collects data on the user's complete answer process for that question. The intelligent learning system records all operations the user performs from the start of answering to submission, including the calculation process at each step, intermediate results, modification history, etc.

[0050] S203. Obtain the standard answer process data of the target question in the preset question bank. The standard answer process data includes the standardized sequence of solution steps and the correct result corresponding to each step. The standard answer process data represents the standardized problem-solving procedures stored in the pre-set question bank, serving as a reference standard for judging the correctness of users' solutions. The standardized problem-solving step sequence refers to the standard operating procedure for completing the question, a set of steps arranged in logical order, with each step having a clear goal and expected output.

[0051] The intelligent learning system executes this step after acquiring the user's incorrect answer process data. Specifically, the system retrieves the standard answer process data for the target question from a pre-set question bank, including the standard solution plan. This standard plan is typically pre-designed by educational experts and includes the optimal solution path and the standard answer for each step. The intelligent learning system parses this standard plan into structured data, including the step sequence, intermediate results, key formulas, etc., providing a benchmark for subsequent error localization and analysis. Simultaneously, the system extracts the importance weight of each step for subsequent judgment of the error's severity.

[0052] Optionally, in some embodiments, the intelligent learning system may employ a dynamic generation strategy. First, the intelligent learning system analyzes the knowledge points and problem-solving requirements of the target question, and retrieves relevant problem-solving rules from the knowledge base. Then, the intelligent learning system automatically derives the optimal problem-solving path based on these rules and calculates the standard result for each step. Finally, the intelligent learning system verifies the generated problem-solving path and results to ensure their correctness and completeness.

[0053] S204. Based on the data of incorrect answering process and the data of standard answering process, determine the characteristics of answering errors, obtain the reasons for errors based on the characteristics of answering errors, generate feedback information based on the reasons for errors, and send the feedback information to the user terminal. The feedback information shall include at least the reasons for errors and corrective suggestions. Among them, error characteristics refer to the key differences and error patterns extracted by comparing the user's actual problem-solving process with the standard problem-solving process. Error causes refer to the specific reasons for the incorrect answer derived from the error characteristics analysis, which may include misunderstandings of concepts, calculation errors, omissions of steps, etc. Feedback information represents the personalized guidance suggestions generated by the intelligent learning system based on the user's error causes, including error location, explanation of the cause, and improvement suggestions.

[0054] The intelligent learning system executes this step after acquiring data on the user's incorrect and standard answer processes. Specifically, the system first meticulously compares the user's answer data with the standard answer data, identifying anomalies in the user's problem-solving process through pattern matching and difference analysis. The system then analyzes the nature and distribution of these anomalies, extracting representative error features. Next, based on pre-defined error diagnosis rules, the system maps these error features to specific error causes. Finally, based on the type of error cause and corrective suggestions, the system generates corresponding feedback information and displays it to the user through their terminal.

[0055] S205. Add the target question to the first preset position in the practice question sequence to update the practice question sequence; Step S205 and Figure 1 The description of step S104 in the above embodiments is similar and will not be repeated here.

[0056] S206. When the answer is correct, obtain the answer process data of the target question. The answer process data includes the answer time, answer step trajectory and historical answer count. Among them, the answer process data refers to a comprehensive set of information that reflects the user's problem-solving behavior and thought process. The answer time represents the total time spent by the user from seeing the question to submitting the final answer. The answer step trajectory refers to a detailed record of the user's problem-solving process, which includes every action the user inputs in chronological order, used to reconstruct the user's complete problem-solving thought process.

[0057] When the intelligent learning system receives the answer returned by the user's terminal and initially determines that it is correct, it executes this step to further analyze the user's true mastery of the knowledge points in the correctly answered questions, distinguishing between different states such as lucky correct answers, barely correct answers, and mastery. Specifically, after the intelligent learning system confirms that the user's answer is correct, it does not immediately end the processing of that question, but instead requests detailed answer process data for that target question from the user's terminal or retrieves it from the server cache.

[0058] S207. Determine the first mastery score based on the ratio of the answering time to the preset standard time for the target question; determine the second mastery score based on the number of historical answers and question types; determine the third mastery score based on the degree of matching between the answering steps and the standard problem-solving steps for the target question. The preset standard time refers to a reference completion time set in advance for each question by educational experts or through statistical analysis of a large number of high-scoring students' answers. This preset standard time can be set by taking the 75th percentile of the answer times of a large number of high-scoring users, or by the curriculum research teacher directly specifying it based on the difficulty and complexity of the question. The first mastery score is a quantitative score representing the user's problem-solving efficiency, reflecting how close the user's answering speed is to the standard speed. The second mastery score is a quantitative score representing the user's familiarity with the question and the stability of their grasp of the knowledge points. The third mastery score is a quantitative score representing the degree to which the user's problem-solving approach conforms to the standard method, measuring the logical rigor and correctness of the user's problem-solving process.

[0059] After successfully acquiring the answer process data for the target question, the intelligent learning system transforms the raw behavioral data into a quantifiable mastery index. Specifically, the intelligent learning system processes and scores three core data points. First, it compares the user's actual answer time with the preset standard time for the question, calculates a ratio, and converts it into a first mastery score ranging from 0 to 100 based on a preset mapping function (e.g., the closer the ratio is to 1, the higher the score; exceeding a certain range, the score decreases). Second, the intelligent learning system checks the historical number of answers and determines a second mastery score based on a decreasing rule; for example, a correct answer on the first attempt earns full marks, a correct answer on the second earns 80 points, and so on. Finally, the intelligent learning system uses algorithms (such as sequence alignment and keyword matching) to compare the user's answer steps with the standard problem-solving steps, and assigns a third mastery score based on the degree of matching.

[0060] S208. The first mastery score, the second mastery score, and the third mastery score are weighted and summed to obtain the comprehensive mastery score. Specifically, the intelligent learning system assigns a preset weighting coefficient to each of the first, second, and third mastery scores. The sum of these weighting coefficients is typically 1, and their value reflects the emphasis the intelligent learning system places on different aspects (efficiency, proficiency, and methodological correctness) when assessing user mastery. The intelligent learning system then performs a weighted sum of the first, second, and third mastery scores based on these weighting coefficients, resulting in the overall mastery score.

[0061] S209. If the overall mastery score is greater than or equal to the preset mastery score, then proceed with the step of removing the target question from the practice question sequence; remove the target question from the practice question sequence. The preset mastery score is used to define whether the user has truly and firmly grasped the knowledge points tested in the question, providing an objective and unified decision-making basis for the intelligent learning system to distinguish between correct answers and mastery. The setting scheme can be based on educational theories and historical data analysis. For example, it can be set to 85 points, which means that the user's overall performance needs to reach a good level or above in order to be recognized as mastering.

[0062] Specifically, after a user correctly answers a target question, the intelligent learning system compares their calculated overall mastery score with a preset mastery score. If the user's score is greater than or equal to this threshold, the intelligent learning system will remove the target question from the practice question sequence.

[0063] S210. If the overall mastery score is less than the preset mastery score and the question type is not a transfer question, then obtain the variant question associated with the target question and add the variant question to the head of the practice question sequence. Specifically, when the intelligent learning system detects that a user's overall mastery score is below a preset threshold, it initiates a reinforcement process. First, the system checks the type of the current target question. If it's determined not to be a transfer question, the system assumes the user's understanding of the knowledge point is unstable and requires immediate variation practice. Therefore, the system uses primary data links to find another related variation question in the question bank and inserts it at the head of the practice question sequence, helping the user immediately practice the same knowledge point from a different angle to deepen understanding and solidify memory. The originally correctly answered target question is removed, but its position is replaced by a related variation question.

[0064] S211. After removing the target question from the practice question sequence, obtain the base score corresponding to each question type from the preset question bank, and obtain the user's answer accuracy rate for each question. The base score of the transfer question is greater than the base score of the variant question, and the base score of the variant question is greater than the base score of the original question. Calculate the user's practice score at the current learning node based on the base score and the answer accuracy rate. The base score for each question type refers to a pre-set score for the original question, its variations, and its application. This score directly reflects the difficulty and value of that question type within the knowledge system. For example, the original question is worth 10 points, the variations 15 points, and the application questions 20 points, reflecting a higher reward for deepening and applying knowledge. The accuracy rate refers to the percentage of questions answered correctly by the user during practice at the current learning node. The practice score refers to the total score accumulated by the user through continuous practice at the current learning node.

[0065] After a question is determined to be fully mastered and successfully removed from the practice question sequence, the intelligent learning system performs this step. Specifically, whenever a user answers a target question correctly, the intelligent learning system updates their practice score for that learning node. First, the intelligent learning system identifies the type of the target question that was just removed (original question, variation question, or transfer question) and finds the corresponding base score from the system configuration. Simultaneously, the intelligent learning system calculates or updates the user's overall accuracy rate so far in real time. Then, the intelligent learning system applies a scoring formula, combining the base score of the question with the user's accuracy rate to calculate the score increment, and adds it to the user's total practice score.

[0066] S212. When the practice score is greater than or equal to the preset passing score and the practice question sequence is not empty, a learning selection prompt is sent to the user terminal. The learning selection prompt includes continuing practice or entering the next learning node. The preset passing score refers to the minimum score threshold used to determine whether a user has reached the minimum mastery requirement for the current learning node, thus qualifying them to proceed to the next stage of learning. This threshold can be determined by teaching experts based on the importance and difficulty of the knowledge point in the course syllabus. For example, the preset passing score for core foundational knowledge nodes could be set at 90 points, while for elective extension nodes it could be set at 75 points. The learning selection prompt is an interactive notification generated by the intelligent learning system on the user's terminal interface, designed to inform the user of the achievement of a stage goal and empower them to choose their subsequent learning path.

[0067] After updating the user's practice score, the intelligent learning system establishes a crucial point where the user gains autonomy in their learning upon reaching the predetermined learning goal. Specifically, whenever the user's practice score is updated, the intelligent learning system immediately performs a dual-condition check: first, it determines whether the user's current practice score is greater than or equal to the preset passing score for that learning node; second, it checks whether there are any remaining questions in the practice question sequence. When both conditions are met, the intelligent learning system pauses the regular question-pushing process and sends a command to the user's terminal, displaying a learning selection prompt in the form of a pop-up window, card, or notification bar. This prompt clearly informs the user, "Congratulations! You have reached the passing standard!" and clearly lists two clickable options: "Continue practicing" and "Proceed to the next learning node." The intelligent learning system then enters a waiting state to receive the user's selection information.

[0068] S213. When the user terminal returns the selection information to continue practicing, keep the state of the practice question sequence unchanged. S214. When the user terminal returns the selection information to proceed to the next learning node, clear the sequence of practice questions.

[0069] Specifically, the intelligent learning system parses the received selection information. When it confirms that the user's intention is to proceed to the next learning node, it first clears the sequence of practice questions. After clearing, according to the logic of the process flow (as described in S106), the intelligent learning system performs a check and finds that the sequence of practice questions is empty, thus triggering the "End Practice" process. In the End Practice process, the intelligent learning system archives data, such as saving the final practice score and generating a learning report, then updates the user's learning map or course progress, and finally guides the user to the learning interface of the next learning node, thereby achieving a smooth transition in the learning process.

[0070] S215. Based on the second data link relationship, obtain the transfer questions associated with the target question; This step specifically includes: Obtain the answer process data, knowledge point data, and a set of candidate transfer questions that have a second data link relationship with the target question for the target question; Extract the question attribute data of each transfer question in the candidate transfer question set. The question attribute data should include at least the difficulty coefficient, the type of ability tested, and the knowledge point data. The degree of knowledge point matching is determined based on the overlap between the knowledge point data of the target question and the knowledge point data of each transfer question; Based on the answer process data, question attribute data, and knowledge point matching degree, the final transfer question is selected from the candidate transfer question set.

[0071] The candidate transfer question set refers to the entire set of transfer questions that are initially screened through the second data linking relationship and are all potentially related to the current target question. Question attribute data is structured information used to describe the inherent characteristics of a question. Difficulty coefficient is one attribute, typically a value between 0 and 1, used to quantify the difficulty level of the question; the type of ability tested is another attribute, indicating which type of thinking ability the question focuses on testing, such as computational problem-solving ability, logical reasoning ability, or spatial imagination ability. Knowledge point matching degree is a calculated value used to represent the degree of overlap between the knowledge point data of the target question and the knowledge point data of each candidate transfer question.

[0072] The intelligent learning system executes this step after the user has firmly grasped a specific knowledge point or question. This occurs in two scenarios: first, when the user's overall mastery score for a target question reaches the required level, and the question is successfully removed from the practice question sequence; second, when the user's total practice score reaches the preset passing score, and the user chooses to continue practicing. Specifically, the intelligent learning system uses the target question the user has just mastered as a reference point to initiate a highly personalized transfer question recommendation process. First, the intelligent learning system acquires the answer process data and knowledge point data for the target question, and based on a second data link relationship, matches a relevant set of candidate transfer questions from a preset question bank. Next, the intelligent learning system analyzes each transfer question in this set, extracting its difficulty coefficient, type of ability tested, and knowledge point data, among other question attribute data. Then, the intelligent learning system performs a matching degree calculation, comparing the knowledge points of the target question with the knowledge points of each candidate transfer question to obtain a quantified knowledge point matching degree. Finally, the intelligent learning system performs a weighted evaluation based on the user's answer process data, the attributes of the candidate transfer questions, and the content relevance between them, thereby selecting the final transfer question that best suits the user's current level and learning status from the set of candidate transfer questions.

[0073] S216. Add the transfer question to the second preset position of the practice question sequence to update the practice question sequence. The transfer question is a question that examines knowledge points related to the target question.

[0074] This step specifically includes: Get the total number of questions in the sequence of practice questions, and calculate the length of the sequence. When the sequence length is greater than or equal to the preset queue length threshold, the addition of migration questions is paused. When the sequence length is less than the preset queue length threshold, the difficulty coefficient of each question in the practice question sequence is obtained, and the average difficulty value of the sequence is calculated. Obtain the estimated time for each question in the practice question sequence, and calculate the sum of the estimated times for each question in the practice question sequence to obtain the total time for the sequence. The estimated time is the baseline time required to complete the questions marked in the preset question bank. The practice load value is obtained by multiplying the total time taken for the sequence by the average difficulty value of the sequence. When the practice load value is greater than or equal to the preset load threshold, the addition of transfer questions will be paused. When the practice load is less than the preset load threshold, execute the step of obtaining the difficulty coefficient of the transfer question to be added; obtain the difficulty coefficient of the transfer question to be added; When the difficulty coefficient of a transfer question is greater than the average difficulty value of the sequence, the transfer question is added to the end of the practice question sequence. When the difficulty coefficient of a transfer question is less than or equal to the average difficulty value of the sequence, the transfer question is added to the first position of the practice question sequence.

[0075] The second preset position refers to a dynamically determined question insertion position based on specific rules, specifically the head or end of the queue in this step. The sequence length value represents the total number of questions in the practice question sequence. The preset queue length threshold is a value used to limit the maximum length of the practice question sequence, preventing the queue from becoming too long and burdening the user; its setting can be determined based on user research (e.g., set to 10). The average difficulty value of the sequence refers to the average difficulty coefficient of all questions in the practice question sequence, used to measure the overall difficulty of the current queue. The estimated time is the baseline time required to complete a single question, as indicated in the question bank. The total sequence time represents the sum of the estimated times for all questions in the practice question sequence. The practice load value is a comprehensive indicator calculated by multiplying the total sequence time by the average sequence difficulty value, used to assess the cognitive load of the user's current practice. The preset load threshold is the upper limit used to control the user's practice load, preventing user fatigue due to excessive workload; its setting can be determined by teaching experts based on the average user's tolerance.

[0076] After selecting the final transfer questions, the intelligent learning system needs to add them to the practice question sequence. Specifically, the intelligent learning system first performs a condition check on the practice question sequence before adding them. The intelligent learning system obtains the sequence length value from the total number of questions. If the sequence length value is greater than or equal to a preset queue length threshold, it determines that the practice question sequence is full and pauses the addition of transfer questions. If the sequence length value has not reached the upper limit, the intelligent learning system obtains the estimated time and difficulty coefficient of each question in the sequence, calculates the total time of the sequence and the average difficulty value of the sequence, and multiplies the two to obtain the practice load value. If the practice load value is greater than or equal to a preset load threshold, it determines that the current load is too high, and the addition of transfer questions is also paused. Only when both the length and load checks pass will the intelligent learning system perform the addition operation. Further, the intelligent learning system obtains the difficulty coefficient of the transfer question to be added and compares it with the average difficulty value of the sequence: if the difficulty coefficient of the transfer question is greater than the average difficulty value of the sequence, the transfer question is added to the end of the practice question sequence; if the difficulty coefficient of the transfer question is less than or equal to the average difficulty value of the sequence, it is added to the beginning of the queue to update the practice question sequence.

[0077] In this embodiment, by employing a closed-loop question-recommendation strategy that deeply analyzes the entire answering process and implements diagnostic feedback, refined mastery assessment, personalized question recommendation, and intelligent practice load control based on the analysis results, the system can accurately discern the user's true level of knowledge mastery, weaknesses, and cognitive state from their question-solving behavior. This effectively solves the problem in related technologies where question-recommendation methods rely solely on the final answer result, leading to coarse recommendation strategies, delayed feedback, and an inability to adapt to individual user differences. Consequently, it achieves a highly personalized and adaptive diagnostic learning path, enabling precise guidance from knowledge consolidation to ability transfer, and improving the user's learning efficiency and depth.

[0078] The intelligent learning system in the embodiments of this invention is described below from the perspective of hardware processing. Please refer to [link / reference]. Figure 3 This is a schematic diagram of the physical device structure of an intelligent learning system in an embodiment of this application.

[0079] It should be noted that, Figure 3 The structure of the intelligent learning system shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0080] like Figure 3As shown, the intelligent learning system includes a CPU 301, which can perform various appropriate actions and processes based on a program stored in the read-only memory ROM 302 or a program loaded from the storage section 308 into the random access memory RAM 303, such as executing the methods described in the above embodiments. The RAM 303 also stores various programs and data required for system operation. The CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An I / O interface 305 is also connected to the bus 304.

[0081] The following components are connected to I / O interface 305: input section 306 including audio input devices, push-button switches, etc.; output section 307 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 308 including a hard disk, etc.; and communication section 309 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 309 performs communication processing via a network such as the Internet. Drive 310 is also connected to I / O interface 305 as needed. Removable media 311, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 310 as needed so that computer programs read from them can be installed into storage section 308 as needed.

[0082] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 309, and / or installed from removable medium 311. When the computer program is executed by CPU 301, it performs the various functions defined in the present invention.

[0083] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0084] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in a flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0085] Specifically, the intelligent learning system of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, it implements a problem-solving method based on an intelligent learning system provided in the above embodiment.

[0086] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the intelligent learning system described in the above embodiments; or it may exist independently and not assembled into the intelligent learning system. The storage medium carries one or more computer programs, which, when executed by a processor of the intelligent learning system, cause the intelligent learning system to implement the problem-solving method based on the intelligent learning system provided in the above embodiments.

[0087] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0088] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0089] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A question pushing method based on an intelligent learning system, characterized in that, The application is applied to an intelligent learning system, comprising: When the learning node of a course displays completion, a parent question corresponding to a knowledge point of the learning node is extracted from a preset question bank, the preset question bank at least including the parent question and a transformed question being associated with each other through a first data link relationship, and a transfer question being associated with the parent question or the transformed question through a second data link relationship; Based on the first data link relationship, the transformed question associated with the parent question is acquired, and the parent question and the transformed question are collectively composed to form an exercise question sequence, the transformed question being a question of the same knowledge point as the parent question but different in form; If the exercise question sequence is not empty, a first question of the exercise question sequence is pushed to a user terminal as a target question, and a user terminal returns an answer result; When the answer result is incorrect, the target question is added to a first preset position of the exercise question sequence to update the exercise question sequence; When the answer result is correct, the target question is removed from the exercise question sequence; based on the second data link relationship, a transfer question associated with the target question is acquired; and the transfer question is added to a second preset position of the exercise question sequence to update the exercise question sequence, the transfer question being a question of a knowledge point associated with the target question; After the exercise question sequence is updated according to the answer result, if the exercise question sequence is empty, the exercise is ended.

2. The method of claim 1, wherein, Before the step of removing the target question from the exercise question sequence, the method further comprises: Acquiring answer process data of the target question, the answer process data including an answer duration, an answer step track and a historical answer number; According to a ratio of the answer duration to a preset standard duration of the target question, a first mastery degree score is determined; according to the historical answer number and a question type, a second mastery degree score is determined; According to a matching degree of the answer step track and a standard solution step of the target question, a third mastery degree score is determined; The first mastery degree score, the second mastery degree score and the third mastery degree score are weighted and summed to obtain a comprehensive mastery degree score; If the comprehensive mastery degree score is greater than or equal to a preset mastery degree score, the step of removing the target question from the exercise question sequence is executed; If the comprehensive mastery degree score is less than the preset mastery degree score, and the question type is not the transfer question, a transformed question associated with the target question is acquired and added to a first position of the exercise question sequence.

3. The method of claim 2, wherein, After the step of removing the target question from the exercise question sequence, the method further comprises: A basic score corresponding to each question type is acquired from a preset question bank, and an answer accuracy of each question of a user is acquired, the basic score of the transfer question being greater than the basic score of the transformed question, and the basic score of the transformed question being greater than the basic score of the parent question According to the basic score and the answer accuracy, an exercise score of the user at a current learning node is calculated; when the exercise score is greater than or equal to a preset passing score and the exercise question sequence is not empty, sending a learning selection prompt to the user terminal, the learning selection prompt including continuing the exercise or entering a next learning node; when the selection information returned by the user terminal is to continue the exercise, maintaining the state of the exercise question sequence unchanged; when the selection information returned by the user terminal is to enter the next learning node, emptying the exercise question sequence.

4. The method of claim 2, wherein, Based on the second data linkage relationship, the migration question associated with the subject title is obtained, specifically including: obtaining the answer process data, knowledge point data of the subject title, and a candidate migration question set having the second data linkage relationship with the subject title; extracting the question property data of each migration question in the candidate migration question set, the question property data at least including a difficulty coefficient, a test ability type and knowledge point data; determining a knowledge point matching degree according to the overlap degree of the knowledge point data of the subject title and the knowledge point data of each migration question; selecting the final migration question from the candidate migration question set according to the answer process data, the question property data and the knowledge point matching degree.

5. The method of claim 4, wherein, adding the migration question to a second preset position of the exercise question sequence to update the exercise question sequence, specifically including: obtaining the total number of questions in the exercise question sequence to calculate a sequence length value; when the sequence length value is greater than or equal to a preset queue length threshold, pausing the addition of the migration question; when the sequence length value is less than the preset queue length threshold, obtaining the difficulty coefficients of the questions in the exercise question sequence to calculate a sequence average difficulty value; obtaining the difficulty coefficient of the migration question to be added; when the difficulty coefficient of the migration question is greater than the sequence average difficulty value, adding the migration question to the end position of the exercise question sequence; when the difficulty coefficient of the migration question is less than or equal to the sequence average difficulty value, adding the migration question to the head position of the exercise question sequence.

6. The method of claim 5, wherein, After the step of obtaining the difficulty coefficients of the questions in the exercise question sequence to calculate a sequence average difficulty value when the sequence length value is less than the preset queue length threshold, the method further includes: obtaining the expected time consumption of each question in the exercise question sequence and calculating the sum of the expected time consumption of each question in the exercise question sequence to obtain a sequence total time consumption, the expected time consumption being the benchmark time required to complete the question marked in a preset question bank; obtaining an exercise load value according to the product of the sequence total time consumption and the sequence average difficulty value; when the exercise load value is greater than or equal to a preset load threshold, pausing the addition of the migration question; when the exercise load value is less than the preset load threshold, performing the step of obtaining the difficulty coefficient of the migration question to be added.

7. The method of claim 1, wherein, Before the step of adding the subject title to the first preset position of the exercise question sequence, the method further includes: obtaining error answer process data of the subject title, the error answer process data including the answer steps and intermediate calculation results of the user; Obtaining standard answer process data of the target question in a preset question bank, the standard answer process data including a standard problem solving step sequence and a correct result corresponding to each step; According to the error answer process data and the standard answer process data, determining an error feature, and obtaining an error cause according to the error feature; According to the error cause, generating feedback information, and sending the feedback information to the user terminal, the feedback information including at least the error cause and a correction suggestion.

8. An intelligent learning system characterized by, The intelligent learning system includes one or more processors and a memory; the memory is coupled with the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors invoke the computer instructions to enable the intelligent learning system to perform the method in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instructions run on the intelligent learning system, the intelligent learning system performs the method in any one of claims 1-7.

10. A computer program product, characterised in that, When the computer program product runs on the intelligent learning system, the intelligent learning system performs the method in any one of claims 1-7.

Images