Knowledge point preorder and postorder relation construction method and system based on user question-doing records

Abstract

The invention provides a knowledge point front-back order relation construction method and system based on user question-doing records, and belongs to the technical field of education data mining, and the method comprises the steps: obtaining a question-doing record data set of each user; dividing the records into learning sessions according to time, creating knowledge point pairs based on knowledge points of the sessions, and distributing time sequence confidence coefficients by counting a time sequence; and determining a logically associated knowledge point pair list by using the textbook catalog and the question-doing records. Calculating a knowledge point mastering state of each user according to an answering result to obtain a knowledge mastering condition probability; for each knowledge point pair, combining the conditional probability and the time sequence confidence coefficient, respectively taking two knowledge points as a preorder to calculate a preorder score and a postorder score, and comparing the preorder score and a preset threshold value to determine a preorder and postorder relationship; and finally, constructing a verification instruction, and generating a pre-order and post-order relation of the target knowledge points by using a large language model verification and correction relation, so that the timeliness and accuracy of relation construction among the knowledge points can be improved, and a basis is provided for personalized learning path planning.

Description

Technical Field

[0001] This application belongs to the field of educational data mining technology, and in particular relates to a method and system for constructing the sequential relationship of knowledge points based on user's question-answering records. Background Technology

[0002] Knowledge point relationship construction is one of the core technologies in the field of smart education. It can identify the sequential dependencies between knowledge topics, providing fundamental support for educational applications such as personalized learning path planning and adaptive test question recommendation. Establishing this relationship is of great significance for building knowledge graphs and optimizing teaching sequences. Especially in large-scale online education platforms, automatically mining the logical progression relationships between knowledge points can significantly improve learning efficiency and teaching quality.

[0003] Existing methods for constructing knowledge point relationships primarily rely on manual annotation, such as manually building knowledge systems based on course outlines or teacher experience. Some studies have attempted to construct knowledge point relationships through association rule mining, for example, by using the Apriori algorithm or collaborative filtering-based recommendation mechanisms to discover relationships between knowledge points. These techniques can reveal the co-occurrence of knowledge points or the similarity of questions to some extent, but they often remain at a static or superficial level.

[0004] However, existing methods generally have shortcomings: manual annotation is time-consuming and labor-intensive, and difficult to adapt to dynamic updates, making it difficult to maintain effectiveness under large-scale learning data; the Apriori algorithm can only identify the co-occurrence of knowledge points, but cannot distinguish their logical pre- and post-order relationships, resulting in a lack of directionality in the constructed associations; while collaborative filtering methods focus too much on the similarity at the question level, ignoring the inherent logical progression between knowledge points. These shortcomings lead to a lack of timeliness and accuracy in constructing relationships between knowledge points, and the learning path planning does not match the actual learning process, failing to provide a basis for personalized learning path planning, thus affecting the effectiveness of personalized learning. Therefore, there is an urgent need for a method that can automatically and accurately mine the causal dependencies between knowledge points from real problem-solving behavior data and construct the pre- and post-order relationships between knowledge points. Summary of the Invention

[0005] The purpose of this application is to provide a method, system, electronic device, and storage medium for constructing the sequential relationship between knowledge points based on user's question-answering records, in order to solve the problem that the construction of relationships between knowledge points in the prior art lacks timeliness and accuracy, and cannot provide a basis for personalized learning path planning.

[0006] To address the aforementioned technical problems, firstly, this application provides a method for constructing the sequential relationship of knowledge points based on user's question-answering records, including: Obtain the question-answering record dataset for each user. The question-answering record dataset includes multiple question-answering records sorted in chronological order. Each question-answering record includes user ID, question ID, knowledge points associated with the question, answer result, and timestamp. The dataset of each user's question-answering records is divided into multiple learning sessions according to time intervals. Multiple sets of knowledge point pairs with sequential order are created based on the related knowledge points in each learning session. By statistically analyzing the time order of question-answering records related to different knowledge points in all learning sessions, a time order confidence score is assigned to each set of knowledge point pairs. Using the pre-set textbook catalog information and a dataset of practice questions, identify logically related knowledge point pairs among all knowledge points to obtain a list of knowledge point pairs; Based on the answer results in the question-answering record dataset, calculate the mastery status of each knowledge point for each user, and determine the conditional probability of mastery of each knowledge point based on the mastery status. For each knowledge point pair in the knowledge point pair list, based on the conditional probability of knowledge mastery and the temporal confidence, a set of preceding and following scores is calculated using the two knowledge points in the knowledge point pair as preceding knowledge points. The preceding and following relationship of the knowledge points in each knowledge point pair is determined by comparing them with the preset scoring threshold. A validation instruction is constructed for the sequential relationship of knowledge points. The validation instruction is processed using a large language model to obtain the validation result. The validation result is then used to correct the sequential relationship of knowledge points and generate the sequential relationship of the target knowledge points.

[0007] In one feasible implementation, using a pre-set textbook catalog information and a dataset of practice questions, logically related knowledge point pairs are identified across all knowledge points, resulting in a list of knowledge point pairs, including: Using the chapter affiliation information of knowledge points in the textbook's table of contents, we filter out knowledge point pairs that have logical relationships from all knowledge points to obtain the first knowledge point pair list; Using the chapter affiliation information of knowledge points, a prior relation matrix of knowledge points is constructed. Each element in the prior relation matrix of knowledge points represents the strength value of the logical association between knowledge point pairs. By using association rule mining algorithms, we identify knowledge point pairs in the question-answering record dataset where the proportion of users learning a certain knowledge point is greater than a preset proportion, and obtain a list of candidate knowledge point pairs. The intensity value is used as the prior weight of the knowledge point pairs in the candidate knowledge point pair list, and the candidate knowledge point pair list is filtered using a preset weight threshold to obtain the second knowledge point pair list. The knowledge point pair list is determined based on the intersection of the first knowledge point pair list and the second knowledge point pair list.

[0008] In one feasible implementation, the dataset of each user's question-answering records is divided into multiple learning sessions according to time intervals. Multiple sets of sequential knowledge point pairs are created based on the related knowledge points in each learning session. By statistically analyzing the temporal order of question-answering records related to different knowledge points across all learning sessions, a temporal confidence score is assigned to each knowledge point pair, including: The dataset of each user's question-answering records is divided into multiple learning sessions according to time intervals. Each learning session includes a set of question-answering records arranged in chronological order. Based on the order in which the knowledge points associated with the questions appear in each learning session, multiple pairs of knowledge points with sequential order are created. Each pair of knowledge points includes a first knowledge point and a second knowledge point, with the first knowledge point preceding the second knowledge point. Based on the timestamp corresponding to the question ID associated with the knowledge point in each learning session, calculate the number of times the first knowledge point appears before the second knowledge point in any pair of knowledge points in the same learning session, and the number of times the first knowledge point and the second knowledge point appear together. In all learning sessions, the temporal confidence of each knowledge point pair is obtained by calculating the ratio of the total number of times the first knowledge point appears before the second knowledge point in each knowledge point pair to the total number of times the first and second knowledge points appear together.

[0009] In one feasible implementation, the question-answering record further includes a resource learning identifier indicating whether there is an associated resource learning behavior when answering a question with a question ID. The resource learning identifier includes the associated knowledge point and the learning timestamp. The method further includes: When the resource learning identifier indicates that there is a related resource learning behavior when answering questions for a question ID, it is determined whether there are common knowledge points between the knowledge points associated with the resource-related knowledge points and the knowledge points associated with the question. When there are common knowledge points between the knowledge points associated with the resource-related knowledge points and the knowledge points associated with the question, the learning timestamp is associated with the common knowledge points. When both the first and second knowledge points of a knowledge point pair have related learning timestamps, and the learning timestamp associated with the first knowledge point is earlier than the learning timestamp associated with the second knowledge point, the time-series confidence of the knowledge point pair is weighted using a preset weighting coefficient to obtain the updated time-series confidence.

[0010] In one feasible implementation, the method further includes: When there are learning timestamps associated with the first and second knowledge points of a set of knowledge point pairs, and the learning timestamp associated with the first knowledge point is earlier than the learning timestamp associated with the second knowledge point, the decay weight is determined in a preset exponential decay function based on the time interval between the learning timestamp associated with the first knowledge point and the learning timestamp associated with the second knowledge point. By using preset weighting coefficients and attenuation weights, the time series confidence scores of knowledge points are weighted to obtain updated time series confidence scores.

[0011] In one feasible implementation, the knowledge mastery conditional probability includes the knowledge point pair mastery conditional probability and the knowledge point pair non-mastery conditional probability. For each knowledge point pair in the knowledge point pair list, based on the conditional probability of knowledge mastery and temporal confidence, a set of pre- and post-preceding scores is calculated using the two knowledge points in the knowledge point pair as preceding knowledge points, including: For the first knowledge point in each knowledge point pair Second knowledge point Calculate the first pre- and post-order scores. Second and previous order scores :

[0012]

[0013] in, For time series confidence, For mastering the knowledge points, conditional probability. This refers to the conditional probability of a knowledge point not yet mastered.

[0014] In one feasible implementation, the sequential relationship of knowledge points in each knowledge point pair is determined by comparing it with a preset scoring threshold, including: If the score of the first preceding and following knowledge point is greater than the score of the second preceding and following knowledge point, and the score of the first preceding and following knowledge point is greater than the preset scoring threshold, the first knowledge point is determined to be the preceding knowledge point of the second knowledge point, thus obtaining the preceding and following knowledge point relationship of the knowledge point pair; If the score of the second preceding and following knowledge point is greater than the score of the first preceding and following knowledge point, and the score of the second preceding and following knowledge point is greater than the preset scoring threshold, the second knowledge point is determined to be the preceding knowledge point of the first knowledge point, and the preceding and following knowledge point relationship of the knowledge point pair is obtained.

[0015] Secondly, this application provides a system for constructing the sequential relationship of knowledge points based on user's question-answering records, including: The acquisition module is used to acquire the question-answering record dataset for each user. The question-answering record dataset includes multiple question-answering records sorted in chronological order. Each question-answering record includes user ID, question ID, knowledge points associated with the question, answer result, and timestamp. The allocation module is used to divide each user's question-answering record dataset into multiple learning sessions according to time intervals, create multiple sets of knowledge point pairs with sequential order based on the related knowledge points in each learning session, and assign temporal confidence to each set of knowledge point pairs by statistically analyzing the temporal order of question-answering records related to different knowledge points in all learning sessions. The identification module is used to identify logically related knowledge point pairs among all knowledge points using a pre-set textbook catalog information and a dataset of practice questions, and to obtain a list of knowledge point pairs. The determination module is also used to calculate the mastery status of each user for each knowledge point based on the answer results in the question-answering record dataset, and to determine the conditional probability of mastery of each knowledge point based on the mastery status. The determination module is also used to calculate a set of preceding and following scores for each knowledge point pair in the knowledge point pair list, based on the conditional probability of knowledge mastery and temporal confidence, taking the two knowledge points in the knowledge point pair as preceding knowledge points respectively, and determine the preceding and following relationship of knowledge points in each knowledge point pair by comparing with the preset scoring threshold. The correction module is used to construct verification instructions for the pre- and post-order relationships of knowledge points, process the verification instructions using a large language model to obtain verification results, and use the verification results to correct the pre- and post-order relationships of knowledge points to generate the pre- and post-order relationships of the target knowledge points.

[0016] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor is used to implement the steps of the method for constructing the sequential relationship of knowledge points based on the user's question-answering records, as described in the first aspect above, when executing a computer program.

[0017] Fourthly, this application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the steps of the method for constructing the sequential relationship of knowledge points based on user's question-answering records as described in the first aspect above.

[0018] The method for constructing the sequential relationship of knowledge points based on user's question-answering records provided in this application automatically constructs the sequential relationship of knowledge points using real question-answering behavior data, effectively overcoming the problems of low efficiency of manual annotation and lack of directionality of static association rules in existing technologies. By dividing question-answering records into learning sessions and calculating temporal confidence, the method can dynamically capture the order between knowledge points during the learning process, ensuring that the relationship has logical progression. By combining textbook table of contents information to filter logically related knowledge point pairs and assessing learning dependence based on the conditional probability of knowledge mastery, the accuracy and timeliness of relationship construction are enhanced. Finally, the relationship is verified and corrected through a large language model, further improving the matching degree between the sequential relationship of knowledge points and the real learning process, thus providing a reliable basis for personalized learning path planning and significantly improving learning efficiency and teaching quality.

[0019] Furthermore, by introducing resource learning identifiers and their timestamps, we can more precisely capture users' active learning behaviors during the learning process, thereby compensating for the potential random biases that may arise from relying solely on the order of question-solving. When both knowledge points and resource learning behaviors are related and have the same time sequence, weighting is used to enhance the confidence of the time sequence, thus improving the reliability of judging the preceding and following relationships. Furthermore, by introducing exponential decay weights based on the resource learning time interval, recent consecutive learning behaviors contribute more significantly to relationship construction, which is more in line with the natural laws of knowledge accumulation and forgetting, thereby significantly improving the accuracy and timeliness of inferring the preceding and following relationships of knowledge points. Ultimately, this method can construct logical relationships of knowledge points that better fit the actual learning process, providing a more solid and reliable basis for personalized learning path planning. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating a method for constructing the sequential relationship of knowledge points based on user's question-answering records, provided in an embodiment of this application; Figure 2 A flowchart illustrating a method for generating a list of knowledge point pairs provided in an embodiment of this application; Figure 3 This application provides a schematic diagram of the structure of a system for constructing the sequential relationship of knowledge points based on user's question-answering records. Figure 4This application provides a schematic diagram of another system for constructing the sequential relationship of knowledge points based on user's question-answering records. Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in one embodiment of this application. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] The core of this application is to provide a method for constructing the sequential relationship of knowledge points based on user's question-answering records. A flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes steps S110 to S160.

[0024] S110: Obtain the question-answering record dataset for each user. The question-answering record dataset includes multiple question-answering records sorted in chronological order. Each question-answering record includes user ID, question ID, knowledge points associated with the question, answer result, and timestamp.

[0025] The question-answering record dataset refers to a collection of data containing user answering behavior information obtained from online learning platforms, including multiple question-answering records arranged chronologically. The question-associated knowledge point refers to the mapping relationship between questions and knowledge points based on subject textbook chapters and examination outlines, supporting association of multiple knowledge points for a single question; for example, a geometry question can be associated with both triangle congruence and the Pythagorean theorem. The answer result refers to the user's response status, including both correct and incorrect answers. The timestamp is the time stamp of the user's answer, accurate to the millisecond level.

[0026] Obtain a dataset of each user's question-answering records, including extracting the mapping relationship between questions and knowledge points from the database. This mapping relationship supports the association of multiple knowledge points for a single question. Clean the acquired question-answering records, filtering out invalid records with a question-answering time of less than 10 seconds to avoid invalid answer data affecting the analysis results. Aggregate multiple answers from the same user to the same question, using multiple answers within a fixed time interval, with the last answer as the final answer record for that user on that question. Sort the cleaned question-answering records by timestamp to ensure the records are arranged chronologically. Based on the mapping relationship between questions and knowledge points, add the associated knowledge points to each question-answering record, forming a complete question-answering record dataset containing user ID, question ID, associated knowledge points, answer result, and timestamp.

[0027] For example, taking junior high school mathematics as an example, the answer records of 1000 junior high school mathematics users are obtained from an online learning platform. User set in This represents the nth student. (User) The problem-solving record is represented as in Indicates the title. Indicates whether the answer is correct or incorrect. Indicates the time allotted for answering questions. This indicates that the answering time is in milliseconds. Retrieves the set of questions from the answering history. The correspondence between knowledge point tags is as follows: The data preprocessing of the answer records includes filtering invalid records with a time limit of less than 10 seconds, and aggregating multiple answers from the same user to the same question, using the last answer as the final result. The cleaned answer records of 1000 users are sorted by timestamp to form a chronologically arranged dataset. Each record contains a user ID, question ID, related knowledge points, answer result, and timestamp. Optionally, it may include a resource learning identifier to indicate whether there was any associated resource learning behavior during the answering of the question, such as watching explanatory videos. This identifier includes the related knowledge points and the corresponding learning timestamp. This provides high-quality foundational data support for subsequent mining of the sequential relationships between knowledge points.

[0028] In one embodiment, during the data acquisition and preprocessing in step S110, the question-answering record may further include an error type label in addition to the user ID, question ID, related knowledge points, answer result, and timestamp. This error type label is used to identify the specific reason for the user's incorrect answer, such as a conceptual misunderstanding, calculation error, or misinterpretation of the question.

[0029] When calculating the mastery status of knowledge points, error type labels can be used to optimize the calculation. The specific implementation is as follows: When an error record's error type label is "Conceptual Misunderstanding," it indicates that the user has a fundamental misunderstanding of the knowledge point related to the question. In this case, when calculating the mastery of that knowledge point, this error record can be given higher weight, or the answer can be directly considered valid evidence of a lack of mastery. Conversely, if the error type label is "Calculation Error" or "Question Misunderstanding," it indicates that the error may stem from accidental operational errors or habitual issues, rather than a lack of knowledge. When processing such records, their weight in the mastery calculation can be reduced, or they can be treated as noise data and appropriately cleaned during data preprocessing. By introducing error type labels, the nature of user errors can be more precisely distinguished, thereby more accurately assessing the user's true mastery of knowledge points and providing a more reliable data foundation for inferring the sequential relationships of subsequent knowledge points.

[0030] S120: Divide each user's question-answering record dataset into multiple learning sessions according to time intervals. Create multiple sets of knowledge point pairs with sequential order based on the related knowledge points in each learning session. By statistically analyzing the time order of question-answering records related to different knowledge points in all learning sessions, assign a temporal confidence score to each set of knowledge point pairs.

[0031] A learning session refers to a set of continuous question-answering behaviors performed by a user within a fixed time interval, including a sequence of question-answering records ordered by timestamps. A knowledge point pair refers to two knowledge points that appear sequentially in a learning session, with the earlier knowledge point being the preceding knowledge point and the later knowledge point being the following knowledge point. Temporal confidence refers to the frequency with which the preceding knowledge point appears before the following knowledge point in a knowledge point pair.

[0032] Each user's practice data is divided into multiple learning sessions at preset time intervals of 30 minutes. Consecutive practice records are grouped into the same learning session based on timestamps. Multiple knowledge point pairs are created based on the order in which the knowledge points associated with the questions appear in each learning session. For example, the sequence of knowledge points appearing in a learning session might be... Then knowledge point pairs are generated. , , The study analyzes the occurrence of each knowledge point pair across all learning sessions, calculating the number of times a preceding knowledge point appears before a subsequent knowledge point within the same knowledge point pair, as well as the total number of times the knowledge point pairs co-occur. Based on the statistical results, the study calculates the temporal confidence score, for example, for knowledge point pairs. It appears 100 times in all learning sessions, among which exist If it has occurred 85 times previously, the time series confidence level is 0.85.

[0033] S130: Using the pre-set textbook catalog information and the dataset of practice questions, identify logically related knowledge point pairs among all knowledge points to obtain a list of knowledge point pairs.

[0034] The knowledge point pair list refers to the set of knowledge point pairs that have logical relationships among all knowledge points. The first knowledge point pair list refers to the set of knowledge point pairs filtered based on the chapter affiliation information of knowledge points in the textbook catalog. The knowledge point prior relation matrix refers to a matrix constructed using the chapter affiliation information of knowledge points, where each element represents the logical relationship strength of the knowledge point pair. The candidate knowledge point pair list refers to the set of knowledge point pairs identified by association rule mining algorithms in the question-answering record dataset where the proportion of users learning a certain knowledge point pair is greater than a preset proportion. The second knowledge point pair list refers to the set of knowledge point pairs filtered using the strength values ​​in the knowledge point prior relation matrix as prior weights and a preset weight threshold.

[0035] Using the chapter affiliation information of knowledge points in the textbook's table of contents, knowledge point pairs with logical relationships are selected to obtain the first knowledge point pair list, CandSet1. For example, knowledge points belonging to the same or adjacent chapters are considered to have a logical relationship. A priori relation matrix of knowledge points is constructed using the chapter affiliation information, where the matrix element values ​​are calculated based on chapter distance; for example, the strength of a similar chapter is 1.0, the strength of adjacent chapters is 0.8, and the strength of non-adjacent chapters is 0. An association rule mining algorithm is used to calculate the proportion of users who learned each knowledge point pair in the question-answering record dataset. If the proportion is greater than a preset proportion, it is added to the candidate knowledge point pair list. Using the strength values ​​in the priori relation matrix as prior weights, a preset weight threshold is used to filter the candidate list, resulting in the second knowledge point pair list, CandSet2. The final knowledge point pair list, CandSet, is determined based on the intersection of the first and second knowledge point pair lists.

[0036] In one embodiment, in step S130, when constructing the prior relation matrix of knowledge points using the textbook catalog information, in addition to calculating the logical association strength based on the chapter affiliation information of the knowledge points, natural language processing technology can also be introduced to conduct in-depth analysis of the textbook content to obtain a more accurate strength value.

[0037] In practice, natural language processing (NLP) analysis can be performed on textual materials such as official teaching outlines, chapter texts of textbooks, and definitions and descriptions of knowledge points. For example, word embedding models (such as Word2Vec and BERT) can be used to calculate the semantic similarity between different knowledge point names or descriptive texts, or topic models (such as LDA) can be used to analyze the topic distribution relevance of knowledge points in the textbook content. Based on the NLP analysis results, semantic association strength can be assigned to knowledge point pairs. For example, if two knowledge points are highly related semantically or frequently appear in the same teaching context, even if they do not belong to the same or adjacent chapters, they can be assigned a higher strength value. Finally, the semantic association strength based on NLP and the association strength based on chapter distance can be weighted and fused to jointly constitute the element values ​​in the prior relation matrix of knowledge points. This approach can capture more subtle logical connections between knowledge points at the semantic level of textbook content, making up for the shortcomings of relying solely on chapter structure, and making the prior relation matrix more reflective of the inherent connections of knowledge.

[0038] S140: Based on the answer results in the question-answering record dataset, calculate the mastery status of each knowledge point for each user, and determine the conditional probability of mastery of each knowledge point based on the mastery status.

[0039] Mastery status refers to the degree of understanding of a knowledge point, with 1 representing mastery and 0 representing no mastery. Mastery is defined as Accuracy(u,k) = (number of questions k answered correctly by user u) / (total number of questions k answered by user u). When Accuracy(u,k) ≥ 80%, the mastery status is recorded as 1; otherwise, it is recorded as 0. The conditional probability of mastering a knowledge point refers to the probability of mastering a knowledge point... Under the condition of being mastered, knowledge points The probability of being controlled, i.e. The knowledge point refers to the conditional probability knowledge point that has not been mastered. Knowledge points that have not been mastered The probability that is not known, i.e. Based on the answer results in the question-answering record dataset, the mastery level of each knowledge point is calculated for each user, and the mastery status is determined according to the 80% threshold.

[0040] For each knowledge point pair Calculate the conditional probability of mastering a knowledge point: ,in, Represents the set of all users; Represents a single user; Indicates user Knowledge points The mastery status is represented by 1 for mastery and 0 for non-mastery; the numerator represents the knowledge points mastered simultaneously. and The number of users; the denominator represents the knowledge points mastered. The number of users.

[0041] Calculate the conditional probability of not mastering a knowledge point: in, Indicates user Knowledge points The state of not mastering; the molecule represents the knowledge points that have not been mastered at the same time. and The number of users, where the denominator represents the number of knowledge points not yet mastered. The number of users.

[0042] For example, among 10,000 students, the knowledge points The number of questions answered correctly is 6,800, and the number of questions completed is 8,000, indicating a high level of mastery. =0.85≥0.8, Mastery status is 1; Knowledge point The number of questions answered correctly is 6,000, and the number of questions completed is 8,000, indicating a high level of mastery. =0.75 < 0.8, mastery level is 0. For knowledge points... At the same time, master and The number of users is 5,200, and they have mastered... The number of users is 7,500, therefore At the same time, it was not mastered. and The number of users is 1,500, and the exact number is unknown. The number of users is 2,500, therefore .

[0043] S150: For each knowledge point pair in the knowledge point pair list, based on the conditional probability of knowledge mastery and the temporal confidence, calculate a set of preceding and following scores using the two knowledge points in the knowledge point pair as preceding knowledge points, and determine the preceding and following relationship of knowledge points in each knowledge point pair by comparing them with the preset scoring threshold.

[0044] The order of scores refers to the score for knowledge points. middle for The scoring criteria for prerequisite knowledge points are calculated using the following formula: .in For time series confidence, For mastering the knowledge points, conditional probability. This represents the conditional probability of not mastering a knowledge point. The preset scoring threshold is 0.5, used to determine the sequential relationship between knowledge points.

[0045] For each knowledge point pair in the knowledge point pair list Based on time series confidence Calculate the first and last order scores for the conditional probability of knowledge mastery. Second and previous order scores Compare the two scores, if and Then determine for Prerequisite knowledge points; if and Then determine for Prerequisite knowledge points.

[0046] S160: Construct verification instructions for the pre- and post-order relationships of knowledge points, process the verification instructions using a large language model to obtain verification results, and use the verification results to correct the pre- and post-order relationships of knowledge points to generate the pre- and post-order relationships of the target knowledge points.

[0047] Verification instructions refer to prompts constructed based on subject area, judgment target, and reasoning basis. For example, in junior high school mathematics, it might analyze the order of precedence between the sum of the interior angles of a triangle and triangle congruence. Verification results refer to the order of precedence relationships returned by the large language model after processing the verification instructions. The order of precedence relationships of target knowledge points refers to the final order of precedence relationships of knowledge points obtained after verification and correction by the large language model, which has higher educational logical rationality.

[0048] Based on the sequential relationships of knowledge points obtained from S150, a verification instruction is constructed that includes the subject area, judgment target, and reasoning basis. For example, for knowledge point pairs... The generated verification instructions are based on junior high school mathematics, and the analysis... and The order of precedence is determined. The large language model is called to process the instruction and obtain the verification result. Based on the verification result, if the large language model determines the precedence relationship to be inconsistent with the result of S150, the precedence relationship of the knowledge points is corrected. For example, if S150 determines... yes The prerequisite knowledge points, but the large language model returns yes The prerequisite knowledge points are then revised to yes Prerequisite knowledge points.

[0049] For example, S150 determines the knowledge point pair The order of precedence is yes The prerequisite knowledge points, among which It is a linear equation in one variable. This is a quadratic equation in one variable. The verification instruction is constructed based on junior high school mathematics, analyzing the order of the linear and quadratic equations. The instruction is processed by a large language model, and the verification result returns that the linear equation should precede the quadratic equation, i.e. yes Therefore, there is no need to modify the order of the knowledge points. If there is an inconsistency, then it needs to be corrected. and The sequential relationships of knowledge points were rearranged and adjusted. These relationships were then validated twice using a large language model, correcting potential discrepancies between S150 data and teaching principles. This ensured that the dependencies between knowledge points aligned with educational logic, providing an accurate basis for personalized learning path planning.

[0050] This application provides a method for generating a list of knowledge point pairs, and a flowchart of one specific implementation is shown below. Figure 2 As shown, the method includes steps S210 to S250.

[0051] In one feasible implementation, step S130: using the preset textbook catalog information and the dataset of answer records, identify logically related knowledge point pairs among all knowledge points to obtain a list of knowledge point pairs, including: S210: Using the chapter affiliation information of knowledge points in the textbook's table of contents, filter out knowledge point pairs that have logical relationships among all knowledge points to obtain the first knowledge point pair list.

[0052] The first knowledge point pair list refers to a set of knowledge point pairs that have logical relationships, filtered based on the chapter affiliation information of the knowledge points in the textbook catalog. This includes knowledge point pairs within the same chapter and knowledge point pairs from adjacent chapters. For example, knowledge points... and When they belong to the same chapter or adjacent chapters, they are considered to have a logical connection.

[0053] Use the chapter order of knowledge points in the textbook's table of contents to filter knowledge point pairs based on their sequential relationship between chapters. For example, knowledge points... Belongs to Chapter A, Knowledge Points If it belongs to chapter B, and chapter B follows chapter A, then... There is a logical connection. Knowledge point pairs within the same chapter have the highest strength, followed by knowledge point pairs in adjacent chapters. Knowledge point pairs in non-adjacent chapters are not considered to have a logical connection, thus obtaining the first list of knowledge point pairs.

[0054] S220: Using the chapter affiliation information of knowledge points, construct a prior relation matrix of knowledge points. Each element in the prior relation matrix of knowledge points represents the strength value of the logical association between knowledge point pairs.

[0055] A knowledge point prior relation matrix is ​​a matrix constructed using the chapter affiliation information of knowledge points, where each element... Indicates knowledge point pairs The logical association strength value is calculated as follows: 1.0 for the same chapter, 0.8 for adjacent chapters, and 0.0 for non-adjacent chapters.

[0056] Based on the chapter affiliation information of each knowledge point, a pair of knowledge points is created for each knowledge point. Calculate the intensity value. For example, knowledge point. and If they belong to the same chapter, then If they belong to adjacent chapters, then If they are not adjacent chapters, then The rows and columns of the matrix correspond to the knowledge point indices, generating a complete prior relation matrix of knowledge points.

[0057] S230: Using an association rule mining algorithm, identify knowledge point pairs in the question-answering record dataset where the proportion of users learning a certain knowledge point is greater than a preset proportion, and obtain a list of candidate knowledge point pairs.

[0058] The candidate knowledge point pair list refers to the set of knowledge point pairs identified in the question-answering record dataset by the association rule mining algorithm, where the proportion of users who have learned a certain knowledge point pair is greater than a preset proportion, which is usually 50%.

[0059] Use association rule mining algorithms (such as Apriori or FP-Growth) to calculate each knowledge point pair The percentage of users learning in the problem-solving record dataset. For example, among 10,000 students, the percentage of students learning simultaneously. and If the percentage of users is greater than 50% of the total number of users, then... Add to the candidate knowledge point pair list.

[0060] S240: Use the intensity value as the prior weight of the knowledge point pairs in the candidate knowledge point pair list, and use the preset weight threshold to filter the candidate knowledge point pair list to obtain the second knowledge point pair list.

[0061] The second knowledge point pair list refers to the set of knowledge point pairs that uses the strength values ​​in the prior relation matrix of knowledge points as prior weights and is filtered using a preset weight threshold, which is usually 0.5.

[0062] The strength values ​​in the prior relation matrix of knowledge points As a weight, knowledge point pairs with an intensity value greater than a preset weight threshold are selected. For example, intensity value Then keep Intensity value If the answer is no, then remove it. This will give you the second list of knowledge point pairs.

[0063] S250: Determine the knowledge point pair list based on the intersection of the first knowledge point pair list and the second knowledge point pair list.

[0064] The knowledge point pair list refers to the final set of knowledge point pairs determined based on the intersection of the first and second knowledge point pair lists. It includes knowledge point pairs that simultaneously satisfy the logical association in the textbook and the co-occurrence condition of learning behavior.

[0065] Calculate the intersection of the first list of knowledge point pairs and the second list of knowledge point pairs to obtain the final list of knowledge point pairs. For example, if the first list of knowledge point pairs contains... , The second knowledge point is that the list contains , The list of knowledge point pairs is as follows and .

[0066] For example, the textbook's table of contents includes Chapter 1, Rational Numbers (knowledge points) , Chapter 2 Linear Equations in One Variable (Key Points) , Chapter 3 Quadratic Equations in One Variable (Key Points) , Using chapter affiliation information, filter out pairs of items from the same chapter. , , and adjacent chapters , This yields a list of the first knowledge point pairs. A priori relation matrix of knowledge points is constructed, along with their strength values. (Same chapter) (Adjacent chapters). Calculate the knowledge point pairs among 10,000 students. The proportion of learning users is 70% > 50%. The proportion is 65% > 50%, resulting in a list of candidate knowledge point pairs. , Filter by intensity value. Strength 1.0 > 0.5 Since the strength is 0.8 > 0.5, the second knowledge point pair list contains... , Finally, by taking the intersection, the final list of knowledge point pairs is as follows: and .

[0067] In one feasible implementation, step S120: Divide each user's question-answering record dataset into multiple learning sessions according to time intervals; create multiple sets of knowledge point pairs with sequential order based on related knowledge points in each learning session; and assign temporal confidence to each knowledge point pair by statistically analyzing the temporal order of question-answering records related to different knowledge points across all learning sessions, including: The dataset of each user's question-answering records is divided into multiple learning sessions based on time intervals. Each learning session includes a set of question-answering records arranged in chronological order.

[0068] Each user's practice questions are divided into multiple learning sessions based on time intervals. For example, with a 30-minute interval, consecutive practice questions completed within that 30-minute period constitute one learning session. If a user continues practicing after 30 minutes, a new learning session begins. For instance, if a user completed 10 questions between 9:00 and 9:30 and 8 questions between 9:35 and 10:00, this constitutes two learning sessions. Each learning session contains practice questions arranged chronologically, representing the order in which the user learned the knowledge points within that session.

[0069] Based on the order in which the knowledge points associated with the questions appear in each learning session, multiple pairs of knowledge points with sequential order are created. Each knowledge point pair includes a first knowledge point and a second knowledge point, with the first knowledge point preceding the second knowledge point.

[0070] A knowledge point pair refers to a combination of knowledge points within the same learning session, arranged according to their order of appearance. Knowledge point pairs are created based on the order in which the knowledge points associated with the questions appear in each learning session. For example, the order in which the knowledge points appear in a learning session is... , , , The knowledge points include , , Only retain the knowledge point pairs that appear before the second knowledge point in the first knowledge point pair. , , .

[0071] Based on the timestamp corresponding to the question ID associated with each knowledge point in each learning session, calculate the number of times the first knowledge point appears before the second knowledge point in any pair of knowledge points in the same learning session, and the number of times the first and second knowledge points appear together.

[0072] Based on the timestamp corresponding to the question ID associated with each knowledge point in each learning session, time-series statistics of knowledge point pairs within the same learning session are calculated. For example, for knowledge point pairs... In the learning session, calculation exist The number of times it appeared before and and The number of times they co-occur. Suppose in a learning session, and They appeared together 5 times, among which exist If it has occurred 4 times previously, then exist The previous occurrence was 4 times, and the number of times they appeared together was 5 times.

[0073] In all learning sessions, the temporal confidence of each knowledge point pair is obtained by calculating the ratio of the total number of times the first knowledge point appears before the second knowledge point in each knowledge point pair to the total number of times the first and second knowledge points appear together.

[0074] The formula for calculating time series confidence is: Time series confidence = ( exist (total number of times it appeared before) / ( and (Total frequency of co-occurrences). Statistics on the number of knowledge points in all learning sessions. middle exist The sum of the previous occurrences and and The ratio of the total number of times they appear together. For example, among 10,000 students, and They appeared together 10,000 times, among which exist If the sequence has occurred 8,500 times previously, the time series confidence level is 0.85. This value indicates... yes Confidence level of prerequisite knowledge points.

[0075] For example, 10,000 students participated in the learning, and their work records were divided into learning sessions at 30-minute intervals. In 100 learning sessions, the knowledge points were... They appeared together 500 times, among which exist It appeared 425 times previously, therefore the time series confidence level is 0.85. (Knowledge point correct) They appeared together 400 times, among which exist It appeared 340 times previously, with a time series confidence level of 0.85.

[0076] In one feasible implementation, the question-answering record also includes a resource learning identifier indicating whether there is an associated resource learning behavior when answering a question with a question ID. The resource learning identifier includes the associated knowledge point and the learning timestamp. The method further includes: Resource learning identifiers indicate whether a user has used supplementary learning resources (such as videos, handouts, etc.) related to the questions during the problem-solving process. They consist of two parts: resource-related knowledge points and a learning timestamp, used to record the association between the learning resources and the questions. Resource-related knowledge points refer to the set of knowledge points involved in the resource learning content, representing the specific knowledge points the user encountered while learning the resource. For example, if a learning video explains the knowledge points of the sum of the interior angles of a triangle and triangle congruence, the corresponding set of resource-related knowledge points would be... .

[0077] When a resource learning identifier indicates that there is a related resource learning behavior when answering a question for a question ID, it is determined whether there are any common knowledge points between the knowledge points associated with the resource-related knowledge points and the knowledge points associated with the question. When there are common knowledge points between the knowledge points associated with the resource-related knowledge points and the knowledge points associated with the question, the learning timestamp is associated with the common knowledge points.

[0078] When resource learning identifiers indicate associated resource learning behaviors, the set of associated knowledge points for resources and the set of associated knowledge points for questions are extracted, and their intersection is calculated to obtain the common knowledge point set. For example, the set of associated knowledge points for resources is... The set of knowledge points related to the question is as follows: Then the set of common knowledge points is Associate the learning timestamp (e.g., 10:00) in the resource learning identifier with common knowledge points. and .

[0079] When both the first and second knowledge points of a knowledge point pair have related learning timestamps, and the learning timestamp associated with the first knowledge point is earlier than the learning timestamp associated with the second knowledge point, the time-series confidence of the knowledge point pair is weighted using a preset weighting coefficient to obtain the updated time-series confidence.

[0080] In the knowledge point pair In the middle, if and They all have related learning timestamps, and The learning timestamp Before the learning timestamp, a preset weighting coefficient is used to weight the time series confidence. For example, for knowledge point pairs... Time series confidence The preset weighting coefficient is 1.2, and the updated time series confidence level is... If the calculated result exceeds 1.0, it is truncated to 1.0 to ensure that the time series confidence level is within a reasonable range.

[0081] For example, students watched instructional videos related to linear equations in one variable while solving problems; the resource-related knowledge points were... ,in This is a method for solving linear equations in one variable. The study time for applying equations is 10:00 AM. The relevant knowledge points for this question are: ( The problem involves a linear equation in two variables. The time allotted for solving the problem is 10:05. Common knowledge points include: Associate the learning timestamp 10:00 with For knowledge points , The learning time stamp is 10:00. The time series confidence score is calculated before 10:05, based on the timestamp of the question-solving process. The weighting coefficient is 1.1, and the updated time series confidence level is 0.825. This updated time series confidence level provides more accurate input data for S140 to calculate the conditional probability of knowledge mastery and S160 to construct verification instructions, effectively improving the educational logic rationality of the sequential relationship between knowledge points.

[0082] The above embodiments can solve the problem of sparsity in question-solving data: some knowledge points may have few questions, but students learn by referring to supplementary learning resources such as lecture notes. This method uses supplementary behavioral data to make up for the shortcomings of simply solving questions in sequence, making the calculation of sequence confidence richer and more robust. In particular, it can establish reliable relationships between knowledge points with uneven question numbers, ensuring the accuracy of the sequential relationship between knowledge points and the rationality of educational logic.

[0083] In one feasible implementation, the method further includes: When both the first and second knowledge points of a knowledge point pair have associated learning timestamps, and the learning timestamp associated with the first knowledge point precedes the learning timestamp associated with the second knowledge point, a decay weight is determined within a preset exponential decay function based on the time interval between the learning timestamps associated with the first and second knowledge points. Using preset weighting coefficients and decay weights, the corresponding time-series confidence scores for each knowledge point pair are weighted to obtain the updated time-series confidence scores.

[0084] The decay weight refers to a time decay factor designed based on the Ebbinghaus forgetting curve principle, reflecting the degree of decay of the influence of the learning interval on the sequential relationship of knowledge points. The calculation formula is as follows: ,in For the preset attenuation coefficient, The time interval (in minutes) is used, and the weights are designed based on the Ebbinghaus forgetting curve, a classic theory in cognitive psychology, which describes the decay of human memory over time. This model has been incorporated into time-series confidence calculations, addressing the problem of mechanical time-series statistics ignoring the laws governing memory decay. (The last sentence appears to be incomplete and possibly refers to a specific knowledge point.) In the middle, if The learning timestamp Previously, and the time interval Larger (e.g., more than a week), then right The supporting effect of this sequence weakens, therefore its contribution to the confidence level of the time series should be reduced.

[0085] When the knowledge points are correct of and They all have related learning timestamps, and The learning timestamp Previously, the time interval was calculated. Based on the Ebbinghaus forgetting curve principle, a preset decay coefficient is used. Substitute into the exponential decay function Calculate the attenuation weight. For example, The learning timestamp is 10:00. The learning timestamp is 10:30, and the time interval is... minutes, then .

[0086] The updated temporal confidence score incorporates the Ebbinghaus forgetting curve principle, representing the confidence level of a knowledge point in enhancing sequential relationships. The calculation formula is as follows: ,in The original time series confidence level. To preset the weighting coefficients, For attenuation weights. Utilizing preset weighting coefficients. and the calculated attenuation weight Confidence of the original time series Weighting is applied. For example, the original time series confidence score. Decay weight ,but If the calculated result exceeds 1.0, it is truncated to 1.0 to ensure that the confidence level is within a reasonable range.

[0087] For example, knowledge point pairs middle The learning timestamp is 10:00. The learning timestamp is 10:30, and the time interval is... Minutes. Based on the Ebbinghaus forgetting curve principle, calculate the decay weight. Original time series confidence level Preset weighting coefficients Updated time series confidence The updated temporal confidence effectively considers the impact of the learning time interval of knowledge points on cognitive dependence, making the discovered sequential relationships closer to real cognitive dependence rather than accidental temporal sequence. It filters out temporal relationships that may be invalid due to excessively long time intervals, providing more accurate input data for S140 to calculate the conditional probability of knowledge mastery and S160 to construct verification instructions, so that the final knowledge graph more accurately reflects the effective knowledge dependency chain.

[0088] In one feasible implementation, the knowledge mastery conditional probability includes the knowledge point pair mastery conditional probability and the knowledge point pair non-mastery conditional probability.

[0089] The conditional probability of mastering a knowledge point refers to the knowledge point Under the condition of being mastered, knowledge points The probability of being controlled, i.e. The knowledge point refers to the conditional probability knowledge point that has not been mastered. Knowledge points that have not been mastered The probability that is not known, i.e. The sequential score refers to the score used to quantify the relationship between knowledge points. and The scoring index for the strength of sequential dependency is calculated using the following formula: ,in For time series confidence, For mastering the knowledge points, conditional probability. This represents the conditional probability of not mastering a knowledge point. By integrating temporal order with mastery status, the preceding and following order scores effectively capture the causal dependence direction between knowledge points, avoiding the directional bias caused by traditional methods that rely solely on co-occurrence frequency. This provides a quantitative basis for subsequently determining the preceding and following order relationships of knowledge points.

[0090] For each knowledge point pair in the knowledge point pair list, based on the conditional probability of knowledge mastery and temporal confidence, a set of pre- and post-preceding scores is calculated using the two knowledge points in the knowledge point pair as preceding knowledge points, including: For the first knowledge point in each knowledge point pair Second knowledge point Calculate the first pre- and post-order scores. Second and previous order scores :

[0091]

[0092] in, For time series confidence, For mastering the knowledge points, conditional probability. This refers to the conditional probability of a knowledge point not yet mastered.

[0093] In one feasible implementation, the sequential relationship of knowledge points in each knowledge point pair is determined by comparing it with a preset scoring threshold, including: If the score of the first preceding and following knowledge point is greater than the score of the second preceding and following knowledge point, and the score of the first preceding and following knowledge point is greater than the preset scoring threshold, the first knowledge point is determined to be the preceding knowledge point of the second knowledge point, thus obtaining the preceding and following knowledge point relationship of the knowledge point pair.

[0094] If the score of the second preceding and following knowledge point is greater than the score of the first preceding and following knowledge point, and the score of the second preceding and following knowledge point is greater than the preset scoring threshold, the second knowledge point is determined to be the preceding knowledge point of the first knowledge point, and the preceding and following knowledge point relationship of the knowledge point pair is obtained.

[0095] First, based on each knowledge point Time series confidence The conditional probabilities of knowledge point pairs, namely the conditional probability of mastering a knowledge point pair and the conditional probability of not mastering a knowledge point pair, are calculated for each knowledge point pair in the list. Calculate the preceding and following scores for both directions: First preceding and following score measure As Dependency strength of prior knowledge points; Second prior- and post-precedence scores measure As Dependency strength of prerequisite knowledge points.

[0096] Then and In comparison, if and Then determine for The prerequisite knowledge points, among which The preset scoring threshold can be 0.5. and Then determine for The preceding knowledge points. This process ensures that the judgment of the preceding and following order simultaneously satisfies the temporal logic and the correlation with the mastery state, thus improving the accuracy of the results.

[0097] For example, using knowledge points For example, It is a linear equation in one variable. It is a quadratic equation in one variable. Time series confidence level. , , And calculate simultaneously , Calculate the first pre- and post-order scores. Second sequential score Comparison results and ,determination for The prerequisite knowledge point is that a linear equation in one variable is a prerequisite knowledge point for a quadratic equation in one variable. This result will enter the large language model verification stage in step S160 to ensure consistency with the logic of junior high school mathematics teaching, that is, the linear equation in one variable comes first, and the quadratic equation in one variable comes later, providing a reliable basis for personalized learning path planning.

[0098] The method for constructing the sequential relationship of knowledge points based on user's question-answering records provided in this application effectively overcomes the problems of low efficiency of manual annotation and lack of directionality of static association rules in the prior art by automatically constructing the sequential relationship of knowledge points using real question-answering behavior data. By dividing the question-answering records into learning sessions and calculating temporal confidence, the method can dynamically capture the order between knowledge points during the learning process, ensuring that the relationship has logical progression. By combining the textbook table of contents information to filter logically related knowledge point pairs and assessing learning dependence based on the conditional probability of knowledge mastery, the accuracy and timeliness of relationship construction are enhanced. Finally, the relationship is verified and corrected through a large language model, further improving the matching degree between the sequential relationship of knowledge points and the real learning process, thereby providing a reliable basis for personalized learning path planning and significantly improving learning efficiency and teaching quality.

[0099] Figure 3This application provides a schematic diagram of a specific implementation of a system for constructing the sequential relationship of knowledge points based on user's question-answering records. (Refer to...) Figure 3 The system may include: The data acquisition and preprocessing module 310 is used to acquire a dataset of each user's question-answering records and to perform data cleaning and preprocessing on the dataset. Specifically, the module extracts users' historical question-answering records from the online learning platform database, including user ID, question ID, answer result, timestamp, and answer duration, while also acquiring the mapping relationship between questions and knowledge points, supporting the association of one question with multiple knowledge points. The acquired question-answering records are cleaned, including filtering invalid records with an answer duration less than a preset threshold (e.g., 10 seconds), and aggregating multiple answer records of the same user on the same question within the same time interval, using the last answer result as the final record. This results in a uniformly formatted, chronologically arranged dataset of question-answering records, providing a high-quality data foundation for subsequent analysis.

[0100] The sequential relationship analysis engine module 320 is used to analyze the sequential relationships of knowledge points in the preprocessed question-answering record dataset. This module first divides each user's question-answering records into multiple learning sessions based on time intervals. Based on the order of knowledge point appearance in each learning session, multiple sets of knowledge point pairs are created, and the temporal frequency of knowledge point pairs in all learning sessions is statistically analyzed, calculating the temporal confidence of each knowledge point pair. Subsequently, a priori relationship matrix of knowledge points is constructed using the chapter affiliation information in the textbook's table of contents. Logically related knowledge point pairs are filtered out, and association rule mining algorithms, such as the Apriori algorithm, are used to mine frequently co-occurring knowledge point pairs from the question-answering records, forming a candidate knowledge point pair list. Based on the answer results in the question-answering records, each user's mastery status for each knowledge point is calculated, and the conditional probability of the knowledge point pair is calculated based on the mastery status, including the conditional probability of mastery and the conditional probability of non-mastery. For each knowledge point pair, combined with the temporal confidence and conditional probability, sequential scores in two directions are calculated separately. By comparing these scores with preset scoring thresholds, the sequential relationship of the knowledge point pair is determined.

[0101] The large-scale model verification module 330 is used to perform secondary verification of the pre- and post-order relationships of knowledge points output by the pre- and post-order relationship analysis engine module 320. This module constructs a verification instruction (Prompt) based on the subject area, judgment target, and reasoning basis, calls the large language model to process the verification instruction, and obtains the verification result. If the pre- and post-order relationships returned by the large language model are inconsistent with the output of the engine module, the relationships are corrected according to the verification result to generate the pre- and post-order relationships of the target knowledge points, ensuring that the relationships conform to educational logic and subject rules.

[0102] The relational graph output module 340 is used to structure and organize the sequential relationships of knowledge points after verification and correction by large models, and then output them to a graph database such as Neo4j for storage. The knowledge point relational graph generated by this module can support educational applications such as personalized learning path planning and adaptive question recommendation, providing a dynamic and accurate foundation for knowledge dependencies in online learning systems.

[0103] Figure 4 This application provides a schematic diagram of the structure of another specific implementation of a system for constructing the sequential relationship of knowledge points based on user's question-answering records, as shown in the following embodiment. Figure 4 The system may include: The acquisition module 410 is used to acquire the question-answering record dataset for each user. The question-answering record dataset includes multiple question-answering records sorted in chronological order. Each question-answering record includes user ID, question ID, knowledge points associated with the question, answer result, and timestamp. The allocation module 420 is used to divide each user's question-answering record dataset into multiple learning sessions according to time intervals, create multiple sets of knowledge point pairs with sequential order based on the related knowledge points in each learning session, and assign temporal confidence to each set of knowledge point pairs by statistically analyzing the temporal order of question-answering records related to different knowledge points in all learning sessions. The determination module 430 is used to identify logically related knowledge point pairs among all knowledge points by using the preset textbook catalog information and the dataset of question-answering records, and to obtain a list of knowledge point pairs; The determination module 430 is also used to calculate the mastery status of each user for each knowledge point based on the answer results in the question-answering record dataset, and to determine the knowledge mastery conditional probability of each knowledge point based on the mastery status. The determination module 430 is also used to calculate a set of preceding and following scores for each knowledge point pair in the knowledge point pair list, based on the conditional probability of knowledge mastery and the temporal confidence, taking the two knowledge points in the knowledge point pair as preceding knowledge points respectively, and determine the preceding and following relationship of knowledge points in each knowledge point pair by comparing with the preset scoring threshold. The correction module 440 is used to construct verification instructions for the pre- and post-order relationships of knowledge points, process the verification instructions using a large language model to obtain verification results, and use the verification results to correct the pre- and post-order relationships of knowledge points to generate the pre- and post-order relationships of target knowledge points.

[0104] In one embodiment, the determining module 430 is specifically used to: utilize the chapter affiliation information of knowledge points in the textbook catalog information to filter out knowledge point pairs with logical relationships from all knowledge points, obtaining a first knowledge point pair list; utilize the chapter affiliation information of knowledge points to construct a knowledge point prior relation matrix, where each element in the knowledge point prior relation matrix represents the strength value of the logical relationship between knowledge point pairs; use an association rule mining algorithm to determine knowledge point pairs in the question-answering record dataset where the proportion of users learning a certain knowledge point pair is greater than a preset proportion, obtaining a candidate knowledge point pair list; use the strength value as the prior weight of the knowledge point pairs in the candidate knowledge point pair list, and use a preset weight threshold to filter the candidate knowledge point pair list, obtaining a second knowledge point pair list; and determine the knowledge point pair list based on the intersection of the first and second knowledge point pair lists.

[0105] In one embodiment, the allocation module 420 is specifically used to divide each user's question-answering record dataset into multiple learning sessions according to time intervals. Each learning session includes a set of question-answering records arranged in chronological order. Based on the order of appearance of the knowledge points associated with the questions in each learning session, multiple sets of knowledge point pairs with sequential order are created. Each knowledge point pair includes a first knowledge point and a second knowledge point, with the first knowledge point appearing before the second knowledge point. Based on the timestamp corresponding to the question ID associated with the knowledge point in each learning session, the number of times the first knowledge point appears before the second knowledge point in any set of knowledge point pairs in the same learning session, and the number of times the first and second knowledge points appear together, are calculated. In all learning sessions, the sum of the number of times the first knowledge point appears before the second knowledge point in each set of knowledge point pairs, and the ratio of this sum to the sum of the number of times the first and second knowledge points appear together, are calculated to obtain the temporal confidence of each set of knowledge point pairs.

[0106] In one embodiment, the question-answering record further includes a resource learning identifier indicating whether there is an associated resource learning behavior when answering a question for a question ID. The resource learning identifier includes a resource-related knowledge point and a learning timestamp. The allocation module 420 is further configured to determine whether the knowledge point associated with the resource-related knowledge point and the question-related knowledge point have the same common knowledge point when the resource learning identifier indicates that there is an associated resource learning behavior when answering a question for a question ID. When the knowledge point associated with the resource-related knowledge point and the question-related knowledge point have the same common knowledge point, the learning timestamp is associated with the common knowledge point. When both the first and second knowledge points of a knowledge point pair have associated learning timestamps, and the learning timestamp associated with the first knowledge point is earlier than the learning timestamp associated with the second knowledge point, the temporal confidence of the knowledge point pair is weighted using a preset weighting coefficient to obtain an updated temporal confidence.

[0107] In one embodiment, the allocation module 420 is further configured to determine a decay weight in a preset exponential decay function based on the time interval between the learning timestamps associated with the first knowledge point and the learning timestamp associated with the second knowledge point when both the first knowledge point and the second knowledge point in a set of knowledge point pairs have associated learning timestamps, and the learning timestamp associated with the first knowledge point is earlier than the learning timestamp associated with the second knowledge point; and to weight the corresponding time series confidence of the knowledge point pairs using preset weighting coefficients and decay weights to obtain updated time series confidence.

[0108] In one embodiment, the conditional probability of knowledge mastery includes the conditional probability of mastering a knowledge point pair and the conditional probability of not mastering a knowledge point pair; the determining module 430 is specifically used to determine the first knowledge point in each knowledge point pair. Second knowledge point Calculate the first pre- and post-order scores. Second and previous order scores :

[0109]

[0110] in, For time series confidence, For mastering the knowledge points, conditional probability. This refers to the conditional probability of a knowledge point not yet mastered.

[0111] In one embodiment, the determining module 430 is specifically used to determine that the first knowledge point is the predecessor of the second knowledge point when the first predecessor score is greater than the second predecessor score and the first predecessor score is greater than a preset scoring threshold, thereby obtaining the predecessor-precedence relationship of the knowledge points in the knowledge point pair; and to determine that the second knowledge point is the predecessor of the first knowledge point when the second predecessor score is greater than the first predecessor score and the second predecessor score is greater than the preset scoring threshold, thereby obtaining the predecessor-precedence relationship of the knowledge points in the knowledge point pair.

[0112] The knowledge point pre- and post-order relationship construction system based on user's question-answering records in this application embodiment is used to implement the aforementioned knowledge point pre- and post-order relationship construction method based on user's question-answering records. Therefore, the specific implementation of the knowledge point pre- and post-order relationship construction system based on user's question-answering records can be found in the embodiment section of the knowledge point pre- and post-order relationship construction method based on user's question-answering records above. The specific implementation can be referred to the description of the corresponding embodiment, and will not be repeated here.

[0113] Figure 5 A schematic diagram of the hardware structure of an electronic device provided in one embodiment of this application is shown.

[0114] The electronic device may include a processor 510 and a memory 520 storing computer program instructions.

[0115] Specifically, the processor 510 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0116] Memory 520 may include mass storage for data or instructions. For example, and not limitingly, memory 520 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 520 may include removable or non-removable (or fixed) media. Where appropriate, memory 520 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 520 is non-volatile solid-state memory.

[0117] Memory may include read-only memory (ROM), random access memory (RAM), disk storage media devices, optical storage media devices, flash memory devices, and electrical, optical, or other physical / tangible memory storage devices. Therefore, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to the first aspect of this disclosure.

[0118] The processor 510 reads and executes computer program instructions stored in the memory 520 to implement any of the above embodiments of the method for constructing the sequential relationship of knowledge points based on the user's question-answering records.

[0119] In one example, the electronic device may also include a communication interface 530 and a bus 540. Wherein, such as Figure 5 As shown, the processor 510, memory 520, and communication interface 530 are connected through bus 540 and complete communication with each other.

[0120] The communication interface 530 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0121] Bus 540 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 540 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0122] This electronic device can execute the method for constructing the sequential relationship of knowledge points based on the user's question-answering records in the embodiments of this application, thereby achieving a combination of Figure 1 and Figure 2 The method described is for constructing the sequential relationship of knowledge points based on the user's question-answering records.

[0123] Furthermore, in conjunction with the method for constructing the sequential relationship of knowledge points based on user's question-answering records in the above embodiments, this application embodiment can provide a computer-readable storage medium for implementation. This computer-readable storage medium stores computer program instructions; when executed by a processor, these computer program instructions implement any of the methods for constructing the sequential relationship of knowledge points based on user's question-answering records in the above embodiments.

[0124] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0125] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0126] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0127] The aspects of this application have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0128] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. The above provides a detailed description of a method, system, electronic device, and storage medium for constructing the sequential relationship of knowledge points based on user's question-answering records. Specific examples have been used in this document to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the scope of protection of this application.

Claims

1. A method for constructing the sequential relationship of knowledge points based on user's question-answering records, characterized in that, The method includes: Obtain a dataset of each user's question-answering records. The dataset includes multiple question-answering records sorted chronologically. Each question-answering record includes a user ID, a question ID, a knowledge point associated with the question, the answer result, and a timestamp. The dataset of each user's question-answering records is divided into multiple learning sessions according to time intervals. Multiple sets of knowledge point pairs with sequential order are created based on the related knowledge points in each learning session. By statistically analyzing the time order of the question-answering records associated with different knowledge points in all learning sessions, a time order confidence score is assigned to each set of knowledge point pairs. Using the pre-set textbook catalog information and the dataset of practice questions, identify logically related knowledge point pairs among all knowledge points to obtain a list of knowledge point pairs; Based on the answer results in the question-answering record dataset, calculate the mastery status of each knowledge point for each user, and determine the conditional probability of mastery of each knowledge point based on the mastery status. For each knowledge point pair in the knowledge point pair list, based on the knowledge mastery conditional probability and the temporal confidence, a set of preceding and following scores is calculated using the two knowledge points in the knowledge point pair as preceding knowledge points, and the preceding and following relationship of the knowledge points in each knowledge point pair is determined by comparing with a preset scoring threshold. A verification instruction is constructed for the pre- and post-order relationship of the knowledge points. The verification instruction is processed using a large language model to obtain the verification result. The verification result is then used to correct the pre- and post-order relationship of the knowledge points to generate the pre- and post-order relationship of the target knowledge points.

2. The method according to claim 1, characterized in that, The method utilizes the preset textbook catalog information and the dataset of practice questions to identify logically related knowledge point pairs among all knowledge points, resulting in a list of knowledge point pairs, including: Using the chapter affiliation information of knowledge points in the textbook catalog, knowledge point pairs with logical relationships are selected from all knowledge points to obtain the first knowledge point pair list; Using the chapter affiliation information of the knowledge points, a priori relation matrix of knowledge points is constructed, where each element in the priori relation matrix represents the strength value of the logical association between knowledge point pairs; Using an association rule mining algorithm, we identify knowledge point pairs in the question-answering record dataset where the proportion of users learning a certain knowledge point is greater than a preset proportion, and obtain a candidate knowledge point pair list. The intensity value is used as the prior weight of the knowledge point pairs in the candidate knowledge point pair list, and the candidate knowledge point pair list is filtered using a preset weight threshold to obtain a second knowledge point pair list. The knowledge point pair list is determined based on the intersection of the first knowledge point pair list and the second knowledge point pair list.

3. The method according to claim 1, characterized in that, The process involves dividing each user's question-answering record dataset into multiple learning sessions based on time intervals, creating multiple sets of knowledge point pairs with sequential order based on the associated knowledge points in each learning session, and assigning temporal confidence scores to each knowledge point pair by statistically analyzing the temporal order of the question-answering records associated with different knowledge points across all learning sessions. The dataset of each user's question-answering records is divided into multiple learning sessions according to time intervals, and each learning session includes a set of question-answering records arranged in chronological order; Based on the order of appearance of the knowledge points associated with the questions in each learning session, multiple sets of knowledge point pairs with sequential order are created. Each knowledge point pair includes a first knowledge point and a second knowledge point, with the first knowledge point preceding the second knowledge point. Based on the timestamp corresponding to the question ID associated with the knowledge point in each learning session, the number of times the first knowledge point appears before the second knowledge point in any pair of knowledge points in the same learning session, and the number of times the first knowledge point and the second knowledge point appear together are calculated. In all the learning sessions, the sum of the number of times the first knowledge point appears before the second knowledge point in each knowledge point pair is calculated, and the ratio of this sum to the sum of the number of times the first knowledge point and the second knowledge point appear together is used to obtain the temporal confidence of each knowledge point pair.

4. The method according to claim 3, characterized in that, The question-answering record also includes a resource learning identifier that indicates whether there is an associated resource learning behavior when answering questions for a question ID. The resource learning identifier includes the associated knowledge points and the learning timestamp. The method further includes: When the resource learning identifier indicates that there is an associated resource learning behavior when answering a question for a question ID, it is determined whether the knowledge point associated with the resource-associated knowledge point and the knowledge point associated with the question have the same common knowledge point. When the knowledge point associated with the resource-associated knowledge point and the knowledge point associated with the question have the same common knowledge point, the learning timestamp is associated with the common knowledge point. When both the first and second knowledge points of a knowledge point pair have associated learning timestamps, and the learning timestamp associated with the first knowledge point is earlier than the learning timestamp associated with the second knowledge point, the time-series confidence of the knowledge point pair is weighted using a preset weighting coefficient to obtain the updated time-series confidence.

5. The method according to claim 4, characterized in that, The method further includes: When both the first and second knowledge points of a knowledge point pair have associated learning timestamps, and the learning timestamp associated with the first knowledge point is before the learning timestamp associated with the second knowledge point, a decay weight is determined in a preset exponential decay function based on the time interval between the learning timestamp associated with the first knowledge point and the learning timestamp associated with the second knowledge point. Using preset weighting coefficients and the attenuation weights, the time series confidence scores corresponding to knowledge points are weighted to obtain the updated time series confidence scores.

6. The method according to claim 1, characterized in that, The conditional probability of knowledge mastery includes the conditional probability of mastering a knowledge point pair and the conditional probability of not mastering a knowledge point pair. For each knowledge point pair in the knowledge point pair list, based on the conditional probability of knowledge mastery and the temporal confidence, a set of pre- and post-order scores is calculated using the two knowledge points in the knowledge point pair as preceding knowledge points, including: For the first knowledge point in each knowledge point pair Second knowledge point Calculate the first pre- and post-order scores. Second and previous order scores : ； ； in, The time series confidence level. The probability of mastering the aforementioned knowledge points is given by condition. The conditional probability of not mastering the knowledge point mentioned above.

7. The method according to claim 6, characterized in that, The step of determining the sequential relationship of knowledge points in each knowledge point pair by comparing with a preset scoring threshold includes: If the first preceding-follower score is greater than the second preceding-follower score, and the first preceding-follower score is greater than the preset scoring threshold, then the first knowledge point is determined to be the preceding knowledge point of the second knowledge point, and the preceding-follower relationship of the knowledge point pair is obtained. If the second preceding-follower score is greater than the first preceding-follower score, and the second preceding-follower score is greater than the preset scoring threshold, then the second knowledge point is determined to be the preceding knowledge point of the first knowledge point, and the preceding-follower relationship of the knowledge point pair is obtained.

8. A system for constructing the sequential relationship of knowledge points based on user's question-answering records, characterized in that, include: The acquisition module is used to acquire a dataset of each user's question-answering records. The dataset includes multiple question-answering records sorted in chronological order. Each question-answering record includes a user ID, a question ID, a knowledge point associated with the question, an answer result, and a timestamp. The allocation module is used to divide the dataset of each user's question-answering records into multiple learning sessions according to time intervals, create multiple sets of knowledge point pairs with sequential order based on the associated knowledge points in each learning session, and assign time sequence confidence to each set of knowledge point pairs by statistically analyzing the time order of the question-answering records associated with different knowledge points in all learning sessions. The determination module is used to identify logically related knowledge point pairs among all knowledge points using the preset textbook catalog information and the dataset of question-answering records, and to obtain a list of knowledge point pairs; The determination module is also used to calculate the mastery status of each knowledge point for each user based on the answer results in the question-answering record dataset, and to determine the knowledge mastery conditional probability of each knowledge point pair based on the mastery status; The determination module is also used to calculate a set of preceding and following scores for each knowledge point pair in the knowledge point pair list, based on the knowledge mastery conditional probability and the temporal confidence, taking the two knowledge points in the knowledge point pair as preceding knowledge points, and determine the preceding and following relationship of knowledge points in each knowledge point pair by comparing with a preset scoring threshold. The correction module is used to construct verification instructions for the pre- and post-order relationships of the knowledge points, process the verification instructions using a large language model to obtain verification results, and use the verification results to correct the pre- and post-order relationships of the knowledge points to generate the pre- and post-order relationships of the target knowledge points.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the method for constructing the sequential relationship of knowledge points based on user's question-answering records as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, enables the method for constructing the sequential relationship of knowledge points based on user's question-answering records as described in any one of claims 1 to 7.

Images