A wrong question automatic arrangement method based on artificial intelligence analysis
By using an AI-based error correction method, which determines the correction method based on the error's characteristic values and status, error correction management is optimized, solving the problems of system lag and poor correction effect, and achieving efficient and reasonable error storage and management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU TIYUAN NETWORK TECHNOLOGY CO LTD
- Filing Date
- 2026-03-31
- Publication Date
- 2026-06-30
AI Technical Summary
Existing methods for organizing incorrect questions are prone to system lag, missing questions, and disordered sorting during high-concurrency data collection, resulting in poor organization results. Furthermore, they do not consider the impact of incorrect question storage methods on analysis, thus reducing organization efficiency.
By using artificial intelligence-based analysis methods, the collection status and processing trigger parameters are determined based on characteristic values such as the accuracy of incorrect question identification, the differentiation of question types, the frequency of errors, and the stability of question types. The collection is then processed either immediately or temporarily, and the management of incorrect questions is optimized by combining the priority of knowledge points, the severity of errors, and the weight of the correction.
It improves the efficiency and rationality of error correction, avoids data loss and incorrect error classification, and enhances the accuracy and efficiency of error management.
Smart Images
Figure CN122309812A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing, and in particular to a method for automatically organizing incorrect questions based on artificial intelligence analysis. Background Technology
[0002] With the development of artificial intelligence and image recognition technology, more and more systems are adopting automatic identification and collection methods to complete the collection of incorrect questions. However, existing methods for organizing incorrect questions often use a uniform organization mode, which leads to problems such as system lag, missing questions, and disordered sorting during high-concurrency collection, resulting in poor results in organizing incorrect questions. Therefore, how to achieve adaptability in organizing incorrect questions and thus improve its efficiency has become a core technical problem that urgently needs to be solved.
[0003] Chinese Patent Publication No. CN113761030A discloses a system for statistical analysis of incorrect answers, including an incorrect answer statistical analysis platform, a student terminal, and a teacher terminal. The incorrect answer statistical analysis platform includes: a learning assessment module that outputs assessment questions to the student terminal; based on the student's answers, incorrect answers are identified and stored; the stored incorrect answers are analyzed using an incorrect answer analysis network model to obtain incorrect answer analysis results; a weakness analysis module analyzes the stored incorrect answer analysis results using a weakness analysis model to obtain weakness output results; finally, a learning plan is pushed to the teacher terminal based on the weakness output results. However, while the above technical solution considers determining the teaching plan based on the incorrect answer analysis results, it does not consider the impact of the actual storage method of incorrect answers on subsequent incorrect answer analysis, reducing the effectiveness of incorrect answer storage and thus affecting the efficiency of incorrect answer organization. Summary of the Invention
[0004] To address this, the present invention provides an automatic error correction method and system based on artificial intelligence analysis, which overcomes the problem in the prior art that it does not consider determining the teaching plan based on the error correction analysis results, but does not consider the impact of the actual storage method of the error correction on subsequent error correction analysis, thereby reducing the effectiveness of error correction storage and thus affecting the efficiency of error correction.
[0005] To achieve the above objectives, this invention provides a method for automatically organizing incorrect questions based on artificial intelligence analysis, comprising: The collection status is determined based on the accuracy of incorrect question identification and the discrimination of question types, and the decision on whether to perform incorrect question feature analysis is based on the collection status. The error characteristic state is determined based on the error frequency characteristic value and the error type stability, and the triggering parameters are sorted into knowledge point priority or error severity according to the error characteristic state; The sorting method is determined as either immediate sorting or temporary collection based on the preset trigger range of the sorting trigger parameters. In response to the first triggering condition, the processing order of temporarily deferred error collection is determined based on the priority sorting coefficient, and the number of temporarily deferred error collection errors processed at one time is determined based on the fluctuation of the priority coefficient. In response to the second triggering condition, for the temporarily deferred collection sequence, determine whether to perform weight correction processing based on the difference in question type protocol; Based on the matching degree of incorrect questions, a target catalog is determined, and the completed target incorrect questions are assigned to the corresponding target catalog.
[0006] Furthermore, for data collection states where the accuracy rate of incorrect question identification is less than or equal to the preset accuracy rate or the question type discrimination is greater than the preset question type discrimination, it is determined that incorrect question feature analysis will be performed.
[0007] Furthermore, for error characteristic states where the error frequency characteristic value is greater than the preset error frequency or the error type stability is less than or equal to the preset stability, the trigger parameters are determined to be the priority of knowledge points. For error characteristic states where the error frequency characteristic value is less than or equal to the preset error frequency and the error type stability is greater than the preset stability, the sorting trigger parameter is determined as the error severity.
[0008] Furthermore, when the sorting trigger parameters are within the first preset trigger range, it is determined that the collection of target incorrect questions will be temporarily suspended; when the sorting trigger parameters are within the second preset trigger range, it is determined that the sorting of target incorrect questions will be performed immediately.
[0009] Furthermore, in response to the first triggering condition, a temporary collection sequence is constructed for the temporarily deferred collection of incorrect questions and they are processed periodically; in a single processing, each incorrect question is processed in real time according to the priority coefficient from largest to smallest, and the number of questions processed in a single processing is negatively correlated with the fluctuation of the priority coefficient. The first triggering condition is to organize the triggering parameters as knowledge point priority.
[0010] Furthermore, the priority sorting coefficient is determined based on the relevance of knowledge points and the frequency of incorrect questions; The priority sorting coefficient is positively correlated with both the relevance of knowledge points and the popularity of incorrect questions.
[0011] Furthermore, when the hierarchical coordination imbalance is greater than the preset imbalance, the knowledge point relevance is determined based on the number of effective association paths; When the hierarchical coordination imbalance is less than or equal to the preset imbalance, the relevance of knowledge points is determined based on the number of related circles.
[0012] Furthermore, in response to the second triggering condition, for questions whose difference in question type protocol is greater than the preset difference, the weight correction process is determined to be executed.
[0013] Furthermore, the weight correction process includes: determining the correction weight based on historical accurate reference coefficients, and performing weight fusion to determine the error severity of incorrect questions; the second triggering condition is to organize the triggering parameter as the error severity.
[0014] Furthermore, the matching degree of incorrect questions is determined based on the urgency of incorrect question analysis and the sufficiency of directory space, and the directory with the highest matching degree of incorrect questions is recorded as the target directory.
[0015] Compared with the prior art, the beneficial effects of the present invention are that the collection status is determined based on the accuracy of incorrect question identification and the question type discrimination, and the incorrect question feature analysis is determined based on the collection status. The accuracy of incorrect question identification represents whether the incorrect question identification result is correct, and the question type discrimination represents the degree of difference between question types. This avoids directly organizing incorrect questions when the questions are not identified in a timely manner, which would lead to incorrect question classification, thereby improving the efficiency of incorrect question organization.
[0016] Furthermore, in this invention, the error characteristic state is determined based on the error frequency feature value and the error type stability, and the sorting trigger parameter is determined as the knowledge point priority or the error severity according to the error characteristic state. The error frequency feature value characterizes the frequency of the error in the question, and the error type stability reflects whether the cause of the question error is fixed. Different sorting trigger parameters are set accordingly, which avoids the fact that a single trigger parameter is difficult to effectively adapt to the actual error situation, thereby reducing the efficiency of error sorting and improving the efficiency of error sorting.
[0017] Furthermore, this invention determines the sorting method as immediate sorting or delayed collection based on the preset trigger range of the sorting trigger parameters. The preset trigger range of the sorting trigger parameters indicates whether it is in a high-concurrency state. For high-concurrency states, delayed collection is used to avoid data loss caused by congestion. For low-concurrency states, immediate sorting is used to ensure the rationality of error handling and improve the efficiency of error sorting.
[0018] Furthermore, in this invention, for the temporarily deferred collection sequence, the weight correction process is determined based on the difference in question type protocol. This avoids the problem that when the question type is rich, a single difference in question type protocol is difficult to effectively adapt to different types of wrong questions, resulting in poor error management. This improves the efficiency of error sorting.
[0019] Furthermore, this invention determines the matching degree of incorrect questions based on the urgency of incorrect question analysis and the sufficiency of directory space, and records the directory with the highest matching degree as the target organization directory. The urgency of incorrect question analysis represents the degree of urgency for incorrect questions to be organized, and the sufficiency of directory space represents the remaining capacity of the directory, thereby ensuring that the utilization efficiency of each directory is maximized, neither idle nor overloaded, thus improving the rationality of the matching degree of incorrect questions, and the directory with the highest matching degree of incorrect questions is recorded as the target organization directory. Attached Figure Description
[0020] Figure 1 This is a flowchart of the automatic error sorting method based on artificial intelligence analysis of the present invention; Figure 2 This is a flowchart illustrating the process of determining whether to perform error feature analysis based on the accuracy of error identification and the discrimination of question types in this invention. Figure 3 This is a flowchart illustrating how the present invention determines the triggering parameters based on error frequency characteristic values and error type stability. Figure 4 This is a flowchart illustrating how the present invention determines the sorting method based on the preset trigger range of the sorting trigger parameters. Detailed Implementation
[0021] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.
[0022] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.
[0023] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.
[0024] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.
[0025] Please see Figures 1 to 4 As shown, this invention provides a method for automatically organizing incorrect questions based on artificial intelligence analysis, including: The collection status is determined based on the accuracy of incorrect question identification and the discrimination of question types, and the decision on whether to perform incorrect question feature analysis is based on the collection status. The error characteristic state is determined based on the error frequency characteristic value and the error type stability, and the triggering parameters are sorted into knowledge point priority or error severity according to the error characteristic state; The sorting method is determined as either immediate sorting or temporary collection based on the preset trigger range of the sorting trigger parameters. In response to the first triggering condition, the processing order of temporarily deferred error collection is determined based on the priority sorting coefficient, and the number of temporarily deferred error collection errors processed at one time is determined based on the fluctuation of the priority coefficient. In response to the second triggering condition, for the temporarily deferred collection sequence, determine whether to perform weight correction processing based on the difference in question type protocol; Based on the matching degree of incorrect questions, a target catalog is determined, and the completed target incorrect questions are assigned to the corresponding target catalog.
[0026] Specifically, for data collection states where the accuracy rate of incorrect question identification is less than or equal to the preset accuracy rate or the question type discrimination is greater than the preset question type discrimination, it is determined that incorrect question feature analysis will be performed.
[0027] Error identification accuracy = number of correctly identified questions / total number of questions collected; Question type discrimination = average Euclidean distance between the feature vector of the current question and the feature vector of the standard question type; It is easy to understand that the higher the user's requirements for the effect of error management, the higher the value of the preset recognition accuracy and the higher the value of the preset question type discrimination. This invention provides a preset value of recognition accuracy and preset question type discrimination. In this invention, the preset recognition accuracy is 0.85 and the preset question type discrimination is 0.6.
[0028] Specifically, for error characteristic states where the error frequency characteristic value is greater than the preset error frequency or the error type stability is less than or equal to the preset stability, the trigger parameters are determined to be the priority of knowledge points. For error characteristic states where the error frequency characteristic value is less than or equal to the preset error frequency and the error type stability is greater than the preset stability, the sorting trigger parameter is determined as the error severity.
[0029] For a single incorrect question, the number of errors for that question in the most recent monitoring period is obtained, and this number of errors is recorded as the error frequency characterization value for that question. The method for confirming the stability of the error type is: Error type stability = 1 / the variance of the error type in a single monitoring period.
[0030] It is easy to understand that the higher the user's requirements for the effectiveness of error management, the higher the value of the preset error frequency and the higher the value of the preset error type stability. This invention provides a preset error frequency and preset error type stability value. In this invention, the preset error frequency is 5, in units of times, and the preset error type stability value is 0.7.
[0031] Specifically, when the collection trigger parameters are within the first preset trigger range, it is determined that the collection of the target incorrect questions will be temporarily suspended; when the collection trigger parameters are within the second preset trigger range, it is determined that the collection of the target incorrect questions will be performed immediately.
[0032] The first preset trigger range is when the organizing trigger parameter is greater than the preset organizing trigger parameter, and the second preset trigger range is when the organizing trigger parameter is less than or equal to the preset organizing trigger parameter; the organizing trigger parameter includes the priority of knowledge points and the degree of harm of errors.
[0033] The value of the preset organization trigger parameter is larger when the user has higher requirements for the effect of error management. This invention provides a preset organization trigger parameter value of 0.6.
[0034] Specifically, in response to the first triggering condition, a temporary collection sequence of incorrect questions will be constructed and periodically processed; in a single processing, each incorrect question will be processed in real time according to the priority coefficient from largest to smallest, and the number of questions processed in a single processing is negatively correlated with the volatility of the priority coefficient. The first triggering condition is to organize the triggering parameters as knowledge point priority.
[0035] Priority coefficient volatility = maximum priority adjustment coefficient within the period - minimum priority adjustment coefficient. Priority coefficient volatility characterizes whether the priority adjustment coefficient is stable within the monitoring period. It is easy to understand that the greater the priority coefficient volatility, the less stable the priority adjustment coefficient is within the monitoring period, which in turn affects the stability of the priority. In this case, the number of processes per time should be reduced to avoid system lag.
[0036] The number of items processed at one time = baseline number × preset priority coefficient fluctuation / priority coefficient fluctuation; the value of the baseline number can be understood as follows: the higher the user's requirements for the effect of error management, the larger the value of the baseline number. This invention provides a value of the baseline number, in which the value of the baseline number is 100, and the unit is questions.
[0037] Specifically, the priority sorting coefficient is determined based on the relevance of knowledge points and the frequency of incorrect questions; The priority sorting coefficient is positively correlated with both the relevance of knowledge points and the popularity of incorrect questions.
[0038] Priority sorting coefficient = knowledge point relevance / preset knowledge point relevance + incorrect question popularity / preset incorrect question popularity; for a single question, the total number of people who answered the question incorrectly in the most recent monitoring period is recorded as the number of people to be analyzed, and the incorrect question popularity = number of people to be analyzed / total number of people who answered the question. The preset values of preset knowledge point relevance and preset incorrect question popularity are easily understood to be such that the higher the user's requirements for the effectiveness of incorrect question management, the larger the preset knowledge point relevance and preset incorrect question popularity values will be. This invention provides a preset value of preset knowledge point relevance and preset incorrect question popularity, in which the preset knowledge point relevance is set to 0.8 and the preset incorrect question popularity is set to 0.59.
[0039] Specifically, when the imbalance in hierarchical coordination is greater than the preset imbalance, the relevance of knowledge points is determined based on the number of effective related paths; When the hierarchical coordination imbalance is less than or equal to the preset imbalance, the relevance of knowledge points is determined based on the number of related circles.
[0040] Hierarchical coordination imbalance = |Average number of associations of first-level knowledge points / Average number of associations of second-level knowledge points - 1|; It is easy to understand that the higher the user's requirements for error management efficiency, the larger the value of the preset hierarchical coordination imbalance. This invention provides a preset hierarchical coordination imbalance value, which is 0.2.
[0041] Specifically, in response to the second triggering condition, for questions whose difference in question type protocol is greater than the preset difference, the weight correction process is determined to be executed.
[0042] Specifically, the weight correction process includes: determining the correction weight based on historical accurate reference coefficients, and performing weight fusion to determine the error severity of incorrect questions; the second triggering condition is to organize the triggering parameter as the error severity.
[0043] Specifically, the matching degree of incorrect questions is determined based on the urgency of incorrect question analysis and the sufficiency of directory space, and the directory with the highest matching degree of incorrect questions is recorded as the target directory.
[0044] Error matching degree = α1 × urgency of error analysis + α2 × sufficiency of directory space; where α1 and α2 are weighting coefficients, α1 + α2 = 1. It is easy to understand that if the urgency of error analysis has a greater impact on error matching degree, then the value of α1 is larger, and if the sufficiency of directory space has a greater impact on error matching degree, then the value of α2 is larger. This invention provides a set of values for α1 and α2, where α1 is 0.5 and α2 is 0.5.
[0045] Urgency of error analysis = frequency of errors × importance coefficient of knowledge point; Adequacy of directory space = (total directory capacity - number of questions already stored) / total directory capacity.
[0046] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.
Claims
1. A wrong question automatic arrangement method based on artificial intelligence analysis, characterized in that, include: The collection status is determined based on the accuracy of incorrect question identification and the discrimination of question types, and the decision on whether to perform incorrect question feature analysis is based on the collection status. The error characteristic state is determined based on the error frequency characteristic value and the error type stability, and the triggering parameters are sorted into knowledge point priority or error severity according to the error characteristic state; The sorting method is determined as either immediate sorting or temporary collection based on the preset trigger range of the sorting trigger parameters. In response to the first triggering condition, the processing order of temporarily deferred error collection is determined based on the priority sorting coefficient, and the number of temporarily deferred error collection errors processed at one time is determined based on the fluctuation of the priority coefficient. In response to the second triggering condition, for the temporarily deferred collection sequence, determine whether to perform weight correction processing based on the difference in question type protocol; Based on the matching degree of incorrect questions, a target catalog is determined, and the completed target incorrect questions are assigned to the corresponding target catalog. 2.The method of claim 1, wherein the method further comprises: If the accuracy rate of incorrect question identification is less than or equal to the preset accuracy rate or the question type discrimination is greater than the preset question type discrimination, then it is determined to perform incorrect question feature analysis. 3.The method of claim 1, wherein the method further comprises: When performing error feature analysis, for error feature states where the error frequency feature value is greater than the preset error frequency feature value or the error type stability is less than or equal to the preset stability, the trigger parameter is determined to be the knowledge point priority. For error characteristic states where the error frequency characteristic value is less than or equal to the preset error frequency and the error type stability is greater than the preset stability, the sorting trigger parameter is determined as the error severity. 4.The method of claim 3, wherein, When the collection trigger parameters are within the first preset trigger range, it is determined that the collection of the target incorrect questions will be temporarily suspended; when the collection trigger parameters are within the second preset trigger range, it is determined that the collection of the target incorrect questions will be performed immediately. 5.The method of claim 4, wherein, In response to the first triggering condition, the temporarily delayed collection of incorrect questions is constructed into a delayed collection sequence and processed periodically; in a single processing, each incorrect question is processed in real time according to the priority coefficient from largest to smallest, and the number of questions processed in a single processing is negatively correlated with the fluctuation of the priority coefficient; The first triggering condition is to organize the triggering parameters as knowledge point priority. 6.The method of claim 5, wherein the method further comprises: The priority sorting coefficient is determined based on the relevance of knowledge points and the frequency of incorrect questions. The priority sorting coefficient is positively correlated with both the relevance of knowledge points and the popularity of incorrect questions. 7.The method of claim 6, wherein the method further comprises: When the imbalance in hierarchical coordination exceeds the preset imbalance, the relevance of knowledge points is determined based on the number of effective association paths. When the hierarchical coordination imbalance is less than or equal to the preset imbalance, the relevance of knowledge points is determined based on the number of related circles. 8.The method of claim 7, wherein the method further comprises: In response to the second triggering condition, for questions whose difference in question type protocol is greater than the preset difference, the weight correction process is determined to be executed.
9. The method for automatically organizing incorrect questions based on artificial intelligence analysis according to claim 8, characterized in that, The weight correction process includes: determining the correction weight based on historical accurate reference coefficients, and performing weight fusion to determine the error severity of incorrect questions; the second triggering condition is to organize the triggering parameter as the error severity.
10. The method for automatically organizing incorrect questions based on artificial intelligence analysis according to claim 9, characterized in that, The matching degree of incorrect questions is determined based on the urgency of incorrect question analysis and the sufficiency of directory space, and the directory with the highest matching degree of incorrect questions is recorded as the target directory.