Title determination method, apparatus, device, medium, and program product

Abstract

The application provides a question determination method, device, equipment, medium and program product. The method comprises: obtaining an ability evaluation value of a user; determining a set of alternative questions matched with the ability evaluation value from a preset question bank; the set of alternative questions comprises a plurality of alternative questions; determining a predicted question solving time length of each alternative question according to the ability evaluation value and question attribute information of each alternative question; determining a set of target questions satisfying a preset constraint condition from the set of alternative questions according to the predicted question solving time length of each alternative question; the preset constraint condition comprises that a total predicted question solving time length of all questions in the set of target questions is within a preset target question solving time length interval. In this way, when recommending questions to the user, a set of questions with difficulty adaptation and controllable total time length can be recommended to the user.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and more specifically, to a method, apparatus, device, medium, and program product for determining a topic. Background Technology

[0002] In education, recommending practice questions to reinforce knowledge is a core teaching element. However, when multiple questions are available, current technologies typically rely on the knowledge points to which the questions belong and their preset difficulty levels to determine which practice questions to choose. This approach ignores the user's actual ability level, resulting in practice questions that are difficult to match with different users' true skill levels and making it difficult to accurately estimate the time required for users to complete the practice questions. Summary of the Invention The purpose of this application is to provide a method, apparatus, device, medium, and program product for determining questions, in order to solve the problems existing in related technologies, such as the difficulty in matching practice questions with the actual ability levels of different users and the difficulty in accurately estimating the time required for users to complete practice questions.

[0003] In a first aspect, embodiments of this application provide a method for determining a topic, including: Obtain the user's ability assessment value; A set of candidate questions matching the ability assessment value is determined from a preset question bank; the set of candidate questions includes multiple candidate questions. Based on the ability assessment value and the question attribute information of each of the candidate questions, the predicted time to answer each of the candidate questions is determined. Based on the predicted time to complete each of the candidate questions, a target question group that meets preset constraints is determined from the candidate question set; the preset constraints include: the total predicted time to complete all questions in the target question group is within a preset target time interval.

[0004] In the above implementation process, a set of candidate questions matching the user's ability assessment score is selected from the question bank. This means selecting candidate questions whose difficulty level is appropriate for the user's ability. Therefore, when recommending questions to users later, it effectively avoids recommending overly difficult questions to weaker users or overly easy questions to more capable users, thus achieving a precise match between question difficulty and user ability, significantly improving learning effectiveness. Furthermore, it fundamentally solves the frustration or boredom caused by a mismatch between question difficulty and the user's actual ability, improving the user's problem-solving experience.

[0005] Simultaneously, by combining the user's ability assessment score with the question attribute information of the candidate questions, the predicted time for each candidate question is determined. Based on this, a target question group whose total predicted time falls within the target range is selected. In this way, the time required for the user to complete the target question group can be accurately estimated, improving the predictability of the learning process.

[0006] Optionally, the ability assessment value includes the user's ability value for each knowledge point, and the question attribute information includes at least one knowledge point to which the candidate question belongs and the basic question-solving time corresponding to the candidate question. Based on the ability assessment value and the question attribute information of each of the candidate questions, the predicted time for answering each candidate question is determined, including: For each of the answer choices: From the aforementioned ability assessment values, determine the target ability value for each knowledge point to which the candidate question belongs; The user's overall ability value for the candidate question is determined based on the target ability value. Based on the comprehensive ability value and the basic time required to answer the corresponding question, the predicted answering time for the question is predicted.

[0007] In the above implementation, since the user's ability assessment value includes the user's ability value for each knowledge point, the user's ability assessment is refined to the knowledge point dimension. At the same time, by combining the specific knowledge point to which the candidate question belongs, the user's comprehensive ability value for the candidate question is determined. This allows for a more accurate determination of the user's predicted time to answer a specific candidate question, avoiding prediction bias caused by ignoring the uneven mastery of knowledge points due to the use of a single overall ability value.

[0008] Optionally, when the candidate questions belong to multiple knowledge points, the user's comprehensive ability value for the candidate questions is determined based on the target ability value, including: Calculate the average value of each of the target capability values; the average value is the comprehensive capability value.

[0009] In the above implementation method, when the candidate questions belong to multiple knowledge points, the average of the target ability values ​​is determined as the comprehensive ability value. The comprehensive ability value obtained in this way can reflect the user's overall mastery level of the multiple knowledge points to which the candidate questions belong, avoiding the excessive influence of an excessively high or low ability value of a certain knowledge point on the prediction results, thereby improving the rationality of the predicted question-solving time.

[0010] Optionally, predicting the user's predicted answering time for the candidate questions based on the user's overall ability score for the candidate questions and the basic answering time corresponding to the candidate questions includes: The input information is fed into a pre-trained question-answering time prediction model to obtain the user's predicted answering time for the candidate questions; The input information includes: the user's comprehensive ability value for the candidate questions and the basic time required to answer the corresponding candidate questions.

[0011] In the above implementation, since a user's overall ability score for the candidate questions reflects their overall mastery of those questions, and the basic time allotted for each candidate question reflects its inherent characteristics, the predicted answering time can be obtained by inputting the user's overall ability score and the basic time allotted for each candidate question into the answering time prediction model. This allows for a more accurate determination of the predicted answering time based on the inherent characteristics of the candidate questions and the user's level of mastery.

[0012] Optionally, the ability assessment value includes the user's overall ability value; based on the ability assessment value and the question attribute information of each of the candidate questions, the predicted time for answering each candidate question is determined, including: For each of the above-mentioned alternative topics: The input information is fed into a pre-trained question-answering time prediction model to obtain the user's predicted answering time for the candidate questions; The input information includes the user's overall ability score and the basic time required to complete the questions.

[0013] In the above implementation, since a user's overall ability score reflects their individual ability, and the base time for answering the candidate questions reflects the inherent characteristics of those questions, the predicted answering time for the user on the candidate questions can be obtained by inputting the user's overall ability score and the base time for answering the candidate questions into the answering time prediction model. This allows for accurate determination of the predicted answering time for the user on the candidate questions based on their inherent characteristics and the user's individual ability.

[0014] Optionally, the ability assessment value includes the user's ability value for each knowledge point; based on the ability assessment value and the question attribute information of each of the candidate questions, the predicted time for answering each candidate question is determined, including: For each of the above-mentioned alternative topics: From the ability assessment values, determine the target ability value for each knowledge point to which the candidate questions belong; The input information is fed into a pre-trained question-answering time prediction model to obtain the user's predicted answering time for the candidate questions; The input information includes the target ability value for each knowledge point to which the candidate questions belong and the basic time required to answer the corresponding candidate questions.

[0015] In the above implementation, since the target ability value for each knowledge point to which the candidate question belongs can characterize the user's mastery of the knowledge point involved in that candidate question, and the basic time required to answer the candidate question can reflect the inherent characteristics of the candidate question, the predicted answering time for the user on the candidate question can be obtained by inputting the target ability value for each knowledge point to which the candidate question belongs and the basic time required to answer the candidate question into the answering time prediction model. This allows for accurate determination of the user's predicted answering time on the candidate question based on its inherent characteristics and the user's personalized ability on specific knowledge points.

[0016] Optionally, the input information may also include the user's historical question-answering data; the historical question-answering data consists of answer records for questions that are within a preset time range from the current time and have the same question type as the candidate questions; the answer records include the time spent by the user answering the questions correctly.

[0017] In the above implementation, by using the user's historical question-solving data as input information, the predicted question-answering time is determined by the input question-answering time prediction model. This can effectively integrate the user's recent question-solving status and the evolution of their proficiency in specific question types, thereby significantly improving the timeliness and accuracy of the question-answering prediction model in predicting question-answering time.

[0018] Optionally, a target question group that meets preset constraints is determined from the candidate question set, including: From the set of candidate questions, determine a plurality of candidate question groups that satisfy the preset constraints; For each set of candidate questions, the target value of that set of candidate questions is determined based on the value of each candidate question in that set. The candidate question group with the highest target value is identified as the target question group.

[0019] In the above implementation, by selecting the question group with the highest target value from multiple candidate question groups that meet preset constraints, the target question group can maximize its value while satisfying the preset constraints. Therefore, when recommending the target question group to the user for practice, the expected learning benefits of the user in that quiz can be maximized, and ineffective practice can be reduced.

[0020] Optionally, the value of each of the candidate questions is determined based on the novelty of each candidate question, the number of times the candidate question is answered by the user, and the importance of the knowledge point corresponding to the candidate question. The value of the candidate questions is positively correlated with the novelty of the candidate questions and the importance of the knowledge points corresponding to the candidate questions; the value of the candidate questions is negatively correlated with the number of times the candidate questions are answered by the user.

[0021] In the above implementation method, the more novel the candidate questions are and the more important the knowledge points they belong to, the higher their value; while the more times the candidate questions have been done by users, the lower their value.

[0022] Based on this, questions that correspond to important knowledge points, have high novelty, and have been answered less frequently by users can be identified as high-value questions and included in the target question group. In this way, when recommending the target question group to users for practice, each practice session allows users to focus on key test points, cutting-edge question types, and knowledge gaps, thereby improving learning efficiency and reducing ineffective practice.

[0023] Optionally, the target value of the alternative question group is determined based on the value of each alternative question in the alternative question group, including: The target value of the alternative question group is determined based on at least one of the following factors: the value of each alternative question in the alternative question group, the number of each alternative question in the alternative question group, and the knowledge point coverage of each alternative question in the alternative question group. The target value is positively correlated with the value of each candidate question, the number of each candidate question in the candidate question group, and the knowledge point coverage of each candidate question in the candidate question group.

[0024] In the above implementation, the target value of the candidate question group is positively correlated with the value of each candidate question in the group, the quantity of each candidate question, and the knowledge point coverage of each candidate question. This allows the selection of high-value candidate question groups with sufficient quantity and broad coverage of multiple knowledge points as the target question group. When recommending the target question group to users for practice, the comprehensiveness and richness of the practice can be maximized, effectively reducing training bias caused by missing knowledge points or insufficient practice questions. This allows users to access and consolidate more knowledge points in a single practice session, thus improving their learning efficiency.

[0025] Optionally, the preset constraint conditions further include: The total number of all questions in the target question group is greater than the first preset number, and / or the number of different knowledge points to which all questions in the target question group belong is greater than the second preset number.

[0026] In the above implementation, the preset constraints further include: the total number of all questions in the target question group is greater than a first preset number, and / or, the number of different knowledge points to which all questions in the target question group belong is greater than a second preset number. In other words, this ensures that the number of questions in the target question group is sufficient and the knowledge coverage is broad. Therefore, when recommending the target question group to users for practice, it maximizes the comprehensiveness and richness of the practice, effectively reducing training bias caused by missing knowledge points or insufficient practice questions. This allows users to access and consolidate more knowledge points in a single practice session, thus improving their learning efficiency.

[0027] Optionally, the question determination method further includes: if, within a preset time period, a target question group that meets the preset constraints cannot be determined from the candidate question set, a pre-stored question group to be recommended is determined as the target question group.

[0028] In the above implementation, even if it is not possible to determine the target question group that meets the preset constraints from the candidate question set in a timely manner, the system can still respond to the user's question recommendation request in a timely manner, determine the target question group and recommend it, thus avoiding the user's long period of ineffective waiting.

[0029] Optionally, the method for determining the question may also include: Obtain the user's actual answering time for the target question set; If the actual answering time is greater than the total predicted answering time, and the first difference between the actual answering time and the total predicted answering time is greater than a preset first threshold, then the ability assessment value is reduced based on the first difference; or, If the actual answering time is less than the total predicted answering time, and the second difference between the total predicted answering time and the actual answering time is greater than a preset second threshold, then the ability assessment value is increased based on the second difference.

[0030] In the above implementation, by obtaining the user's actual answering time for the target question set, and adjusting the user's ability assessment value according to the difference between the total predicted answering time and the actual answering time, the ability assessment value can be updated. This allows the predicted answering time determined based on the updated ability assessment value to better reflect the user's true ability.

[0031] Secondly, embodiments of this application also provide a question determining device, comprising: The first acquisition module is used to acquire the user's ability assessment value; The first determining module is used to determine a set of candidate questions that match the ability assessment value from a preset question bank; the set of candidate questions includes multiple candidate questions. The second determining module is used to determine the predicted time to answer each of the candidate questions based on the ability assessment value and the question attribute information of each of the candidate questions. The third determining module is used to determine a target question group that meets preset constraints from the candidate question set based on the predicted time for each candidate question; the preset constraints include: the total predicted time for all questions in the question group is within a preset target time interval.

[0032] Thirdly, embodiments of this application also provide an electronic device, including a processor, a memory, and a communication bus; the communication bus is used to realize the connection and communication between the processor and the memory; the processor is used to execute one or more programs stored in the memory to implement any of the above-mentioned problem determination methods.

[0033] Fourthly, this application also provides a storage medium storing computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions cause the processor to implement the above-mentioned problem-determination method.

[0034] Fifthly, embodiments of this application also provide a computer program product, the computer program product including a computer program, which, when executed by a processor, implements the above-described problem-solving method. Attached Figure Description

[0035] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 A flowchart illustrating a topic determination method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a topic determination device provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0037] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0038] Example 1: To address the problems of existing technologies, such as the difficulty in matching practice questions to the actual ability levels of different users and the difficulty in accurately estimating the time required for users to complete practice questions, this application provides a method for determining practice questions. See also... Figure 1 As shown, Figure 1 This is a flowchart illustrating the topic determination method provided in the embodiments of this application, including: Step S101: Obtain the user's ability assessment value.

[0039] In one embodiment of this application, a user can log in to a terminal device. After verifying the user's identity, the terminal device can establish a session corresponding to the user and obtain the user's user identifier. The user identifier is a unique code used to identify the user. Subsequently, the terminal device can send a question recommendation request to the server. The question recommendation request includes the user identifier. Upon receiving the question recommendation request, the server can perform a matching operation between the user identifier and the ability assessment value to obtain the user's ability assessment value.

[0040] A user's ability assessment value can be pre-set. In some embodiments, when setting an ability assessment value for a new user for the first time, different or default initial ability assessment values ​​can be assigned to all new users. A new user represents a user who has not answered any questions. After recommending questions to users, the ability assessment value for each user can be adjusted based on each user's answer performance, thus obtaining an ability assessment value that is appropriate to each user's true ability.

[0041] Step S102: Determine a set of candidate questions that match the ability assessment value from a preset question bank; the set of candidate questions includes multiple candidate questions. The preset question bank can store multiple candidate questions and the difficulty level corresponding to each candidate question.

[0042] The difficulty level of each question in the alternative question set is higher than the baseline difficulty level, and the difference between the two does not exceed a preset threshold. The baseline difficulty level is determined based on the user's ability assessment. This ensures that the difficulty level of each question in the alternative question set is higher than the user's current ability level, but avoids the problem of the user being unable to answer due to excessive difficulty. Thus, a balance is achieved between challenge and feasibility, effectively guiding learning.

[0043] A user's ability assessment includes their overall ability score. For example, the user's overall ability score can be used as a baseline difficulty level. Alternatively, the baseline difficulty level can be obtained by matching the user's overall ability score against a pre-defined correspondence between overall ability scores and difficulty levels.

[0044] Step S103: Based on the ability assessment value and the question attribute information of each candidate question, determine the predicted time to answer each candidate question.

[0045] Optionally, the predicted time for each candidate question can be determined based on a preset question-answering time prediction model, combined with the ability assessment value and the question attribute information of each candidate question.

[0046] A user's ability assessment score can include their overall ability score and their ability score for each knowledge point. Question attribute information can include at least one knowledge point to which the candidate questions belong and the basic time allotted for each candidate question.

[0047] In this embodiment of the application, for each candidate question, the input information can be determined based on the ability assessment value and the question attribute information of each candidate question. The input information is then input into the answer time prediction model to obtain the predicted answer time for that candidate question.

[0048] The answer time prediction model can be obtained in the following way: Input information with labels is obtained, where the labels are the actual time spent by the user to complete the candidate questions. The input information with labels is then fed into a deep learning model or a decision tree model for training until the loss function value between the output information of the deep learning model or the decision tree model and the label information converges to a preset threshold, thus obtaining the answer time prediction model.

[0049] In one optional implementation of this application, the input information may include the user's overall ability and the basic time required to complete the questions.

[0050] In another optional implementation of this application, the input information may include the user's target ability value for each knowledge point to which the candidate questions belong, and the basic time required to answer the questions.

[0051] In another optional implementation of this application, the input information may include the user's comprehensive ability value for the candidate questions and the basic time required to answer the candidate questions.

[0052] The overall ability score is determined based on the target ability score for each knowledge point to which the candidate questions belong. For example, if there are multiple knowledge points to which the candidate questions belong, the average of the target ability scores can be calculated; the average is the overall ability score. Alternatively, the median of multiple target ability scores can also be used as the overall ability score.

[0053] Based on any of the above optional implementation methods, the input information may also include the user's historical question-answering data; historical question-answering data refers to the answer records of questions that are within a preset time range from the current time and have the same question type as the candidate questions; the answer records include the time spent by the user answering questions correctly. By incorporating the user's historical answer time for the most recent questions of the same type, the answer time prediction model can predict the time based on the user's current state, thereby improving the prediction accuracy.

[0054] Step S104: Based on the predicted time to complete each candidate question, determine the target question group that meets the preset constraints from the candidate question set; the preset constraints include: the total predicted time to complete all questions in the question group is within the preset target time range.

[0055] In this embodiment, if the predicted time for each candidate question is greater than the maximum value of the target time interval, then the candidate question with the shortest predicted time can be selected as the target question group. In other words, in this case, the question with the shortest single-question duration can be identified as a question in the target question group to ensure that questions can be recommended to the user, thereby enabling the user to obtain questions while adhering as closely as possible to the time constraint.

[0056] Normally, the target question group is determined in real time by the server according to the question determination method provided in this application embodiment when it needs to recommend questions to the user. However, in this case, if the target question group cannot be determined from the question bank within a preset time period, the server can directly determine the pre-stored question group to be recommended as the target question group. Correspondingly, the pre-stored question group to be recommended is determined and stored according to the question determination method provided in this application embodiment before it is necessary to recommend questions to the user.

[0057] For example, upon receiving a question recommendation request, the server can run the question determination method provided in this application to generate a target question group. However, if the computation takes too long, preventing the generation of the target question group within a preset time period, the server can call a pre-stored question group to be recommended as the target question group. This pre-stored question group is actually a question group pre-determined by the server during idle periods or upon the last receipt of a question recommendation request, according to the question determination method provided in this application. Thus, even in extreme cases of computational constraints or data anomalies, the server can respond promptly to user question recommendation requests, determine the target question group, and recommend it, avoiding prolonged and ineffective waiting for the user.

[0058] In one optional implementation of this application, the preset constraints may further include: the total number of all questions in the target question group is greater than a first preset number, and / or, the number of different knowledge points to which all questions in the target question group belong is greater than a second preset number.

[0059] However, in specific scenarios where question bank resources are limited, it may be impossible to identify a target question set that satisfies all constraints from the question bank. Therefore, the priority of each constraint can be determined, and questions can be filtered from the question bank in ascending order of priority. Finally, the question set with the highest number of satisfied constraints can be identified as the target question set.

[0060] For example, an initial screening can be performed in a preset question bank based on the highest priority constraint to obtain a set of candidate question groups. This set of candidate question groups may include at least one candidate question group. Subsequently, for each constraint with a priority other than the highest priority, filtering is performed layer by layer using the candidate question groups obtained in the previous round as a benchmark, in descending order of priority. During this process, if the result of filtering for a certain priority constraint is empty, the filtering can be terminated, and the target question group can be determined from the non-empty candidate question group set in the previous round. For example, if there is only one candidate question group in the non-empty candidate question group set in the previous round, that candidate question group can be directly determined as the target question group. If there are multiple candidate question groups in the non-empty candidate question group set in the previous round, any one of the candidate question groups can be determined as the target question group.

[0061] If there are multiple candidate question groups in the non-empty candidate question group set in the previous round. Alternatively, after determining multiple candidate question groups that meet the preset constraints from the candidate question group set, for each candidate question group: the target value of the candidate question group can be determined based on the value of each candidate question in the candidate question group; the candidate question group with the highest target value is determined as the target question group.

[0062] For each set of candidate questions, the target value of the set of candidate questions can be determined based on at least one of the following factors: the value of each candidate question in the set of candidate questions, the number of each candidate question in the set of candidate questions, and the knowledge point coverage of each candidate question in the set of candidate questions.

[0063] Among them, the target value of the alternative question group is positively correlated with the value of each alternative question in the alternative question group, the number of each alternative question in the alternative question group, and the knowledge point coverage of each alternative question in the alternative question group.

[0064] The value of each candidate question can be determined based on its novelty, the number of times it has been answered by users, and the importance of the corresponding knowledge point. Specifically, the value of a candidate question is positively correlated with its novelty and the importance of the corresponding knowledge point; however, it is negatively correlated with the number of times it has been answered by users. The novelty, importance, and number of times each candidate question has been answered can be obtained from its attribute information.

[0065] For example, the value of each candidate question is related to its novelty, the number of times a candidate question is answered by a user, and the importance of the knowledge point corresponding to the candidate question, which in turn satisfies the following formula: Value = w1 × Importance of knowledge point + w2 × Novelty of question - w3 × Number of times to answer; where w1, w2 and w3 are preset weight coefficients.

[0066] If the candidate questions belong to multiple knowledge points, the importance of the corresponding knowledge point can be the average of the importance of the multiple knowledge points to which the candidate questions belong. Alternatively, the importance of the corresponding knowledge point can be the median of the importance of the multiple knowledge points to which the candidate questions belong.

[0067] The question selection method provided in this application selects a set of candidate questions from a question bank that match the user's ability assessment value, i.e., it selects candidate questions whose difficulty level is suitable for the user's ability. This effectively avoids recommending overly difficult questions to users with weaker abilities or overly easy questions to users with higher abilities when recommending questions to them, thus achieving a precise match between question difficulty and user ability and significantly improving learning effectiveness. Furthermore, it fundamentally solves the frustration or boredom caused by a mismatch between question difficulty and the user's actual ability, improving the user's question-solving experience.

[0068] Simultaneously, by combining the user's ability assessment score with the question attribute information of the candidate questions, the predicted time for each candidate question is determined. Based on this, a target question group whose total predicted time falls within the target range is selected. In this way, the time required for the user to complete the target question group can be accurately estimated, improving the predictability of the learning process.

[0069] After step S104, i.e., after determining the target question group, the target question group can be sent to the terminal device where the user is logged in, so that the terminal device can display each question in the target question group. In this way, the user can browse each question in the target question group through the interactive interface of the terminal device and answer the questions. Correspondingly, the terminal device can collect the actual answering time for the target question group, i.e., the total time spent answering all questions in the target question group correctly. The server then feeds back the user's actual answering time for the target question group. After obtaining the user's actual answering time for the target question group, if the actual answering time is greater than the total predicted answering time, and the first difference between the actual answering time and the total predicted answering time is greater than a preset first threshold, the server can lower the ability assessment value based on the first difference; or, if the actual answering time is less than the total predicted answering time, and the second difference between the total predicted answering time and the actual answering time is greater than a preset second threshold, the server can raise the ability assessment value based on the second difference.

[0070] In one optional implementation of this application, if the actual answering time is greater than the total predicted answering time, and the first difference between the actual answering time and the total predicted answering time is greater than a preset first threshold, the server can directly reduce the user's ability assessment value by a preset assessment value.

[0071] Similarly, if the actual answering time is less than the total predicted answering time, and the second difference between the total predicted answering time and the actual answering time is greater than the preset second threshold, the server can directly increase the preset assessment value of the user's ability assessment value.

[0072] In another optional implementation of this application, if the actual answering time is greater than the total predicted answering time, and the first difference between the actual answering time and the total predicted answering time is greater than a preset first threshold, the first difference can be divided by the first threshold and then multiplied by a preset evaluation value to obtain a target reduction value, thereby reducing the user's ability evaluation value by the target reduction value.

[0073] Similarly, if the actual answering time is less than the total predicted answering time, and the second difference between the total predicted answering time and the actual answering time is greater than the preset second threshold, the second difference can be divided by the second threshold and then multiplied by the preset evaluation value to obtain the target improvement value, thereby increasing the user's ability evaluation value by the target improvement value.

[0074] In this way, the user's ability assessment value can be corrected based on the user's actual answer performance, so that the user's ability assessment value can be more in line with the user's true ability. This will allow for the determination of a more accurate predicted answering time based on the updated ability assessment value.

[0075] Example 2: This embodiment, based on Embodiment 1, provides further examples and explanations of the title determination method provided in this application.

[0076] The server-side can include basic data assessment units. These units can store the ability assessment value for each student, i.e., each user. They can also store candidate questions and their attribute information, which may include the difficulty level of the question, the knowledge point it belongs to, and the basic time allotted for completing it.

[0077] The server-side can include a personalized time prediction unit. This unit converts the basic time allotted for each question into a personalized predicted time for each student.

[0078] The personalized time prediction unit includes a question-answering time prediction model. This model is trained and used for prediction employs methods such as gradient boosting decision trees or deep learning.

[0079] For a specific student-option question combination, the input information for the answer time prediction model can include the student's ability assessment value stored in the basic data evaluation unit, the student's historical answer data, and the question attribute information of the option questions. Based on this input information, the answer time prediction model can output the predicted answer time for that student on that option question.

[0080] The server-side can also include an intelligent topic selection and planning unit. This unit is used to identify target topic groups from the candidate topics stored in the basic data evaluation unit.

[0081] First, the intelligent topic selection and planning unit can obtain students' ability assessment values ​​from the basic data assessment unit, and then select a batch of candidate topics of moderate difficulty that match the students' ability assessment values ​​from the candidate topics stored in the basic data assessment unit to form a candidate topic set.

[0082] Then, the personalized time prediction unit can be used to determine the predicted time for students to answer each question in the question set.

[0083] Finally, based on the preset time limit for each candidate question in the candidate question set, the value of each candidate question can be used to determine the target question set that meets the preset constraints. The total predicted time limit for all questions in the target question set is within the preset target time limit range, and the sum of the values ​​of all questions is maximized.

[0084] For example, a preset target time range for answering questions can be used as the knapsack capacity, such as 25 to 35 minutes. Each candidate question in the candidate question set is considered an item, the predicted time for answering questions for each candidate question in the candidate question set is considered the weight of the item, and the value of each candidate question in the candidate question set is considered the value of the item. In this way, at least one group of candidate questions can be obtained whose total weight (corresponding to the total predicted time for answering questions) is within the knapsack capacity (corresponding to the target time range for answering questions).

[0085] For each candidate question set, the target value of the question set is determined based on at least one of the following factors: the value of each question in the question set, the number of questions in the question set, and the knowledge point coverage of each question in the question set. The target value is positively correlated with the value of each question, the number of questions in the question set, and the knowledge point coverage of each question in the question set. The question set with the highest target value is then designated as the target question set.

[0086] If the intelligent question selection and planning unit is unable to determine the target question group within a preset time limit, the intelligent question selection and planning unit can select candidate questions from the candidate question set in order of predicted question-answering time from low to high to form the target question group, thereby ensuring that a target question group can be provided to the user.

[0087] The server side may also include closed-loop feedback and capability update units.

[0088] The closed-loop feedback and ability update unit is used to obtain the actual time a user spends completing the target question set. It then compares and analyzes this actual time with the total predicted time for the target question set, dynamically updating the student's ability assessment value stored in the basic data assessment unit accordingly. For example, if the actual time spent completing the questions is lower than the total predicted time, the student's ability assessment value can be increased.

[0089] In this way, the updated competency assessment can be applied to the next question selection process, enabling the selected set of candidate questions to be adaptively adjusted in real time and become increasingly accurate.

[0090] In addition, the server can generate and store at least one set of questions to be recommended for each student during idle periods or daily, based on the aforementioned question determination method. When responding to a user's question recommendation request, the server can run the aforementioned question determination method in real time to determine the target question set. If the determination fails, the set of questions to be recommended can be directly designated as the target question set, thus providing a final guarantee for the user's experience.

[0091] Example 3: Based on the same inventive concept, this application also provides a topic determination device 200. Please refer to... Figure 2 As shown, Figure 2 It shows the use of Figure 1 The method illustrated describes a problem-determining apparatus 200. It should be understood that the specific functions of the problem-determining apparatus 200 are described above; to avoid repetition, detailed descriptions are omitted here. The problem-determining apparatus 200 includes at least one software function module that can be stored in memory or embedded in the operating system of the problem-determining apparatus 200 in the form of software or firmware. Specifically: See Figure 2 As shown, the question determination device 200 is applied to a server and includes: a first acquisition module 201, a first determination module 202, a second determination module 203, and a third determination module 204.

[0092] The first acquisition module 201 is used to acquire the user's ability assessment value.

[0093] The first determining module 202 is used to determine a set of candidate questions that match the ability assessment value from a preset question bank; the set of candidate questions includes multiple candidate questions.

[0094] The second determining module 203 is used to determine the predicted time to answer each candidate question based on the ability assessment value and the question attribute information of each candidate question.

[0095] The third determining module 204 is used to determine the target question group that meets the preset constraints from the candidate question set based on the predicted time for each candidate question. The preset constraints include: the total predicted time for all questions in the question group is within the preset target time interval.

[0096] In this embodiment of the application, the ability assessment value includes the user's ability value for each knowledge point, and the question attribute information includes at least one knowledge point to which the candidate question belongs and the basic question-solving time corresponding to the candidate question. The second determining module 203 is specifically used for each candidate question: Determine the target ability value for each knowledge point to which the candidate question belongs from the ability assessment value; The user's overall ability value for the candidate question is determined based on the target ability value; Based on the comprehensive ability value and the basic time required to answer the corresponding question, predict the predicted answering time for the question.

[0097] In one feasible embodiment of this application, when there are multiple knowledge points to which the candidate questions belong, the second determining module 203 is specifically used to calculate the average value of each target ability value; the average value is the comprehensive ability value.

[0098] In one feasible implementation of this application, the second determining module 203 is specifically used to input the input information into a pre-trained question-answering time prediction model to obtain the user's predicted question-answering time for the candidate questions; the input information includes: the user's comprehensive ability value for the candidate questions and the basic question-answering time corresponding to the candidate questions.

[0099] In one feasible implementation of this application, the ability assessment value includes the user's overall ability value; the second determining module 203 is specifically used for each candidate question: inputting the input information into a pre-trained answer time prediction model to obtain the user's predicted answer time for the candidate question; the input information includes the user's overall ability value and the basic answer time corresponding to the candidate question.

[0100] In this embodiment of the application, the ability assessment value includes the user's ability value for each knowledge point; the second determining module 203 is specifically used for each candidate question: determining the target ability value for each knowledge point to which the candidate question belongs from the ability assessment value; inputting the input information into a pre-trained answer time prediction model to obtain the user's predicted answer time for the candidate question; the input information includes the target ability value for each knowledge point to which the candidate question belongs and the basic answer time corresponding to the candidate question.

[0101] In one optional implementation of this application, the input information may further include the user's historical question-answering data; the historical question-answering data consists of answer records for questions that are within a preset time range from the current time and have the same question type as the candidate questions; the answer records include the time spent by the user answering questions correctly.

[0102] In this embodiment of the application, the third determining module 204 is specifically used to determine multiple candidate question groups that meet preset constraints from the candidate question set; for each candidate question group, determine the target value of the candidate question group based on the value of each candidate question in the candidate question group; and determine the candidate question group with the highest target value as the target question group.

[0103] In one optional implementation of this application, the value of each candidate question is determined based on the novelty of each candidate question, the number of times the candidate question is answered by the user, and the importance of the knowledge point corresponding to the candidate question; wherein, the value of the candidate question is positively correlated with the novelty of the candidate question and the importance of the knowledge point corresponding to the candidate question; and the value of the candidate question is negatively correlated with the number of times the candidate question is answered by the user.

[0104] In one optional implementation of this application, the third determining module 204 is specifically used to determine the target value of the candidate question group based on at least one of the following factors: the value of each candidate question in the candidate question group, the number of each candidate question in the candidate question group, and the knowledge point coverage of each candidate question in the candidate question group; wherein, the target value is positively correlated with the value of each candidate question, the number of each candidate question in the candidate question group, and the knowledge point coverage of each candidate question in the candidate question group.

[0105] In one optional implementation of this application, the preset constraints may further include: the total number of all questions in the target question group is greater than a first preset number, and / or, the number of different knowledge points to which all questions in the target question group belong is greater than a second preset number.

[0106] In an optional implementation of this application, the third determining module 204 may also be used to determine the pre-stored question group to be recommended as the target question group if, within a preset determined time period, a target question group that meets the preset constraints cannot be determined from the candidate question set.

[0107] In this embodiment of the application, the first acquisition module 201 can also be used to acquire the user's actual answering time for the target question set; if the actual answering time is greater than the total predicted answering time, and the first difference between the actual answering time and the total predicted answering time is greater than a preset first threshold, then the ability assessment value is reduced according to the first difference; or, if the actual answering time is less than the total predicted answering time, and the second difference between the total predicted answering time and the actual answering time is greater than a preset second threshold, then the ability assessment value is increased according to the second difference.

[0108] It should be understood that, for the sake of brevity, some of the content described in Embodiment 1 will not be repeated in this embodiment.

[0109] Example 4: Based on the same inventive concept, this embodiment provides an electronic device, see [link to relevant documentation]. Figure 3 As shown, it includes a processor 301 and a memory 302. Wherein: The processor 301 is used to execute one or more programs stored in the memory 302 to implement the above-mentioned problem-solving method.

[0110] It is understandable that processor 301 can be a processor core or processor chip, or other circuitry capable of program configuration and execution. Memory 302 can be RAM (Random Access Memory), ROM (Read-Only Memory), flash memory, etc., but this is not a limitation.

[0111] It's understandable. Figure 3 The structure shown is for illustrative purposes only; the electronic device may also include components that are more advanced than those shown. Figure 3 The more or fewer components shown, or having the same Figure 3 Different configurations are shown. For example, it may also have an internal communication bus for communication between the processor 301 and the memory 302; or it may have an external communication interface, such as a USB (Universal Serial Bus) interface, a CAN (Controller Area Network) bus interface, etc.; or it may have an information display component such as a display screen, but this is not a limitation.

[0112] Based on the same inventive concept, this embodiment also provides a computer-readable storage medium, such as a floppy disk, optical disk, hard disk, flash memory, USB flash drive, SD (Secure Digital Memory Card), MMC (Multimedia Card), etc., in which one or more programs implementing the above steps are stored. These one or more programs can be executed by one or more processors to implement the above-described method for determining the problem. Further details will not be elaborated here.

[0113] Based on the same inventive concept, this embodiment also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described problem-solving method.

[0114] For example, a computer program product may be an installation package or a program package.

[0115] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0116] Furthermore, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0117] Furthermore, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0118] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0119] In this article, "multiple" refers to two or more.

[0120] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for determining the topic, characterized in that, include: Obtain the user's ability assessment value; A set of candidate questions matching the ability assessment value is determined from a preset question bank; The set of alternative questions includes multiple alternative questions; Based on the ability assessment value and the question attribute information of each of the candidate questions, the predicted time to answer each of the candidate questions is determined. Based on the predicted time required to complete each of the candidate questions, a target question set that meets the preset constraints is determined from the candidate question set. The preset constraints include: the total predicted time for answering all questions in the target question group is within the preset target time interval.

2. The method according to claim 1, characterized in that, The ability assessment value includes the user's ability value for each knowledge point, and the question attribute information includes at least one knowledge point to which the candidate question belongs and the basic question-solving time corresponding to the candidate question. Based on the ability assessment value and the question attribute information of each of the candidate questions, the predicted time for answering each candidate question is determined, including: For each of the answer choices: From the aforementioned ability assessment values, determine the target ability value for each knowledge point to which the candidate question belongs; The user's overall ability value for the candidate question is determined based on the target ability value. Based on the comprehensive ability value and the basic time required to answer the corresponding question, the predicted answering time for the question is predicted.

3. The method according to claim 2, characterized in that, Based on the user's overall ability score for the candidate questions and the basic time required to answer the corresponding candidate questions, the predicted answering time for the user on the candidate questions is predicted, including: The input information is fed into a pre-trained question-answering time prediction model to obtain the user's predicted answering time for the candidate questions; The input information includes: the user's comprehensive ability value for the candidate questions and the basic time required to answer the corresponding candidate questions.

4. The method according to claim 1, characterized in that, The ability assessment value includes the user's ability value for each knowledge point; Based on the ability assessment value and the question attribute information of each of the candidate questions, the predicted time for answering each candidate question is determined, including: For each of the above-mentioned alternative topics: From the ability assessment values, determine the target ability value for each knowledge point to which the candidate questions belong; The input information is fed into a pre-trained question-answering time prediction model to obtain the user's predicted answering time for the candidate questions; The input information includes the target ability value for each knowledge point to which the candidate questions belong and the basic time required to answer the corresponding candidate questions.

5. The method according to any one of claims 1 to 4, characterized in that, From the candidate question set, a target question group that meets the preset constraints is determined, including: From the set of candidate questions, determine a plurality of candidate question groups that satisfy the preset constraints; For each set of candidate questions, the target value of that set of candidate questions is determined based on the value of each candidate question in that set. The candidate question group with the highest target value is identified as the target question group.

6. The method according to claims 1 to 4, characterized in that, Also includes: Obtain the user's actual answering time for the target question set; If the actual answering time is greater than the total predicted answering time, and the first difference between the actual answering time and the total predicted answering time is greater than a preset first threshold, then the ability assessment value is reduced based on the first difference. or, If the actual answering time is less than the total predicted answering time, and the second difference between the total predicted answering time and the actual answering time is greater than a preset second threshold, then the ability assessment value is increased based on the second difference.

7. A problem-determining device, characterized in that, include: The first acquisition module is used to acquire the user's ability assessment value; The first determining module is used to determine a set of candidate questions that match the ability assessment value from a preset question bank; the set of candidate questions includes multiple candidate questions. The second determining module is used to determine the predicted time to answer each of the candidate questions based on the ability assessment value and the question attribute information of each of the candidate questions. The third determining module is used to determine the target question group that meets the preset constraints from the candidate question set based on the predicted time to answer each candidate question. The preset constraints include: the total predicted time for answering all questions in the question group is within the preset target time range.

8. An electronic device, characterized in that, It includes a processor and a memory, the memory storing computer-executable instructions that can be executed by the processor, the processor executing the computer-executable instructions to implement the title determination method according to any one of claims 1 to 6.

9. A storage medium, characterized in that, The storage medium stores computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement the question determination method according to any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the question determination method according to any one of claims 1 to 6.

Images