Learning ability evaluation model optimization method, storage medium and electronic equipment

By introducing data steady-state judgment and correlation evaluation into the learning ability evaluation model, the model parameters are optimized, and the evaluation result deviation problem caused by low data instability and correlation in the prior art is solved, and the accuracy of the model is improved. Effectiveness provides solid data support for personalized education.

CN118917494AInactive Publication Date: 2024-11-08浙江海亮科技有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411398133.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-09
Publication Date
2024-11-08
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing learning ability assessment models are susceptible to data instability and uncorrelation in complex and variable educational data, resulting in bias in optimization results and affecting the accuracy and referenceability of assessment results.

Method used

Before constructing a learning ability assessment model, the steady state of student performance data sets is judged and the correlation between the learning performance prediction results and the learning performance prediction results. If the preset conditions are met, the model parameters are optimized. to ensure data stability and relevance, thereby improving the accuracy of the model.

Benefits of technology

By introducing the correlation judgment of data steady-state judgment and model output results, the inaccurate parameter optimization problems caused by data instability and low correlation are avoided, and the accuracy and effectiveness of the learning ability evaluation model is improved, and it is for schools and parents to Provide more comprehensive and accurate learning tutoring programs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917494A_ABST
    Figure CN118917494A_ABST
Patent Text Reader

Abstract

The invention provides an optimization method of a learning ability evaluation model, a storage medium and electronic equipment. The method comprises the following steps: acquiring a learning ability evaluation model of a student; respectively determining whether the plurality of student score data sets and the plurality of student performance data sets are in a steady state; if the plurality of student score data sets and the plurality of student performance data sets are in a steady state, determining a correlation between an academic score prediction result output by the academic score prediction model and an academic performance prediction result output by the academic performance prediction model; and if the correlation meets a preset condition, optimizing the parameters of the learning ability evaluation model to obtain a target model, thereby ensuring the stability and correlation degree of the data before the parameter optimization process of the learning ability evaluation model, avoiding the problem of inaccurate parameter optimization caused by unstable data and relatively low correlation degree between the data, and improving the accuracy of parameter optimization. And the accuracy and effectiveness of the learning ability evaluation model are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a storage medium and electronic device for optimizing a learning ability assessment model. Background Art

[0002] In recent years, with the deep integration of computer technology and educational informatization, artificial intelligence has been increasingly used in the field of education, especially in the assessment of learning ability. An accurate and scientific learning ability assessment model can provide students with personalized learning guidance and help them adjust their learning strategies. It can also provide valuable references for schools and parents to develop learning plans that better meet students' needs.

[0003] At present, the student learning ability assessment model in related technologies directly uses historical learning performance for prediction. The prediction results obtained by this solution are too objective and do not take into account the factors affecting the student's learning performance. The prediction results are of little reference value and cannot be used to customize learning strategies based on the prediction results. In other solutions, the optimization process based on the student learning ability assessment model generally adopts the method of setting weights or dynamically adjusting model parameters. Although this method can improve the performance of the model to a certain extent, it can only be optimized based on the same model and does not take into account the impact of learning performance on the learning ability assessment results, resulting in low accuracy of the optimized model.

[0004] It can be seen that when the optimization method of the learning ability assessment model in the relevant technology is used for complex and changeable educational data, due to the diversity and instability of educational data, the learning ability assessment model is easily affected by significant differences such as information gap, instability, and irrelevance between data, resulting in deviations in the model optimization results, thereby affecting the accuracy and reference value of the evaluation results. Summary of the invention

[0005] The present disclosure provides an optimization method, storage medium and electronic device for a learning ability evaluation model.

[0006] According to a first aspect of the present disclosure, a method for optimizing a learning ability evaluation model is provided, the method comprising: Obtaining a student's learning ability assessment model, the learning ability assessment model includes a learning performance prediction model and a learning performance prediction model, the learning performance prediction model is established based on multiple student performance data sets, and the learning performance prediction model is established based on multiple student performance data sets; determining whether the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state, respectively; If the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state, determining a correlation between a learning achievement prediction result output by the learning achievement prediction model and a learning performance prediction result output by the learning performance prediction model; If the correlation meets the preset conditions, the parameters of the learning ability evaluation model are optimized to obtain the target model.

[0007] In some embodiments of the present disclosure, determining whether the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state respectively comprises: Respectively obtain a preset time interval and a preset threshold value corresponding to each of the multiple student score data sets and the multiple student performance data sets; For each data set, if the rate of change of each data set in the corresponding preset time interval is less than or equal to the corresponding preset threshold, it is determined that the data set is in a steady state; If the number of data sets in a steady state among the plurality of student achievement data sets and the plurality of student performance data sets meets a preset number condition, it is determined that the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state.

[0008] In some embodiments of the present disclosure, determining the correlation between the learning performance prediction result output by the learning performance prediction model and the learning performance prediction result output by the learning performance prediction model includes: Based on multiple student performance data sets, using the learning performance prediction model, determining multiple learning performance prediction results; Based on multiple student performance data sets, using a learning performance prediction model, determining multiple learning performance prediction results; Respectively determine the score changes and statistical values ​​of the score changes for a plurality of learning score prediction results within a preset time interval, and the performance changes and statistical values ​​of the performance changes for a plurality of learning performance prediction results; Correlation is determined based on the change in grades, the statistical value of the change in grades, the change in performance, and the statistical value of the change in performance.

[0009] In some embodiments of the present disclosure, determining the correlation based on the score change amount, the statistical value of the score change amount, the performance change amount, and the statistical value of the performance change amount includes: The correlation is determined according to the following formula 1: Formula 1 in, Indicates relevance, Indicates the change in performance. It represents the average value of the change in grades. Indicates the amount of change in performance, represents the average value of the performance variation, Indicates the amount of data.

[0010] In some embodiments of the present disclosure, if the correlation meets the preset conditions, the parameters of the learning ability evaluation model are optimized to obtain the target model, including: If the correlation is greater than zero, it is determined that the learning performance prediction result output by the learning performance prediction model is positively correlated with the learning performance prediction result output by the learning performance prediction model; Taking multiple student score data sets and multiple student performance data sets as data samples, and obtaining data labels corresponding to the data samples, the data labels are actual learning ability scores of each student in the data samples; Establishing a regression model, the regression model includes a parameter matrix of a learning ability assessment model; Use data samples and data labels to train the regression model until the training is completed, and obtain the optimized parameter matrix of the learning ability evaluation model; The learning ability evaluation model is updated using the optimized parameter matrix to obtain the target model.

[0011] In some embodiments of the present disclosure, establishing a regression model includes: The regression model is established as shown in the following formula 2: Formula 2 in, represents the data label, , represents the actual learning ability score of the nth student, represents a data sample, represents the parameter matrix, and the elements in the parameter matrix are the partial regression coefficients in the learning achievement prediction model and the learning performance prediction model. Indicates the error value.

[0012] In some embodiments of the present disclosure, the regression model is trained using data samples and data labels until the training is completed, and the optimization parameter matrix of the learning ability evaluation model is obtained, including: The loss function is established as follows: Formula 3 in, represents the loss value, represents the actual learning ability score of the mth student, represents the predicted learning ability score of the mth student calculated using the learning ability assessment model; The regression model is trained using data samples and data labels until the loss value meets the preset loss condition. The training ends and the optimized parameter matrix of the learning ability evaluation model is obtained.

[0013] In some embodiments of the present disclosure, the regression model is trained using data samples and data labels until the loss value meets the preset loss condition, and the training is terminated. The optimization parameter matrix of the learning ability evaluation model includes: During the training process, the parameter matrix of the learning ability evaluation model is expressed as the following formula 4: Formula 4; The optimization parameter matrix is ​​expressed as the following formula 5: Formula 5 in represents the learning rate, Represents the partial derivative of the loss value G; Using the gradient descent method, when the loss value meets the preset loss condition, the parameter matrix corresponding to the loss value that meets the preset loss condition is determined as the optimized parameter matrix.

[0014] According to a second aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0015] According to a third aspect of the present disclosure, an electronic device is provided, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the method described in the first solution when executing the computer program.

[0016] The optimization method, storage medium and electronic device of the learning ability evaluation model provided by the present invention obtain the learning ability evaluation model of the students, the learning ability evaluation model includes a learning performance prediction model and a learning performance prediction model, the learning performance prediction model is established according to multiple student performance data sets, and the learning performance prediction model is established according to multiple student performance data sets; respectively determine whether the multiple student performance data sets and the multiple student performance data sets are in a steady state; if the multiple student performance data sets and the multiple student performance data sets are in a steady state, determine the correlation between the learning performance prediction results output by the learning performance prediction model and the learning performance prediction results output by the learning performance prediction model; if the correlation meets the preset conditions, optimize the parameters of the learning ability evaluation model to obtain the target model, and realize the introduction of steady-state judgment of the data and correlation judgment of the model output results before the parameter optimization process of the learning ability evaluation model to ensure the stability and correlation of the data, avoid the problem of inaccurate parameter optimization due to data instability and low correlation between data, improve the accuracy and effectiveness of the learning ability evaluation model, provide schools and parents with more comprehensive and accurate student learning ability evaluation services, thereby helping schools and parents to formulate and implement effective learning tutoring programs.

[0017] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure. Figure 1 A flowchart of a method for optimizing a learning ability evaluation model provided by an embodiment of the present disclosure; Figure 2 A flowchart of another optimization method for a learning ability evaluation model provided by an embodiment of the present disclosure; Figure 3 A flowchart of another optimization method for a learning ability evaluation model provided by an embodiment of the present disclosure; Figure 4 A flowchart of another optimization method for a learning ability evaluation model provided by an embodiment of the present disclosure; Figure 5 A schematic diagram of the structure of an optimization device for a learning ability evaluation model provided by an embodiment of the present disclosure; Figure 6 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0019] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0020] In recent years, with the deep integration of computer technology and educational informatization, artificial intelligence has been increasingly used in the field of education, especially in the assessment of learning ability. An accurate and scientific learning ability assessment model can provide students with personalized learning guidance and help them adjust their learning strategies. It can also provide valuable references for schools and parents to develop learning plans that better meet students' needs.

[0021] At present, the student learning ability assessment model in related technologies directly uses historical learning performance for prediction. The prediction results obtained by this solution are too objective and do not take into account the factors affecting the student's learning performance. The prediction results are of little reference value and cannot be used to customize learning strategies based on the prediction results. In other solutions, the optimization process based on the student learning ability assessment model generally adopts the method of setting weights or dynamically adjusting model parameters. Although this method can improve the performance of the model to a certain extent, it can only be optimized based on the same model and does not take into account the impact of learning performance on the learning ability assessment results, resulting in low accuracy of the optimized model.

[0022] It can be seen that when the optimization method of the learning ability assessment model in the relevant technology is used for complex and changeable educational data, due to the diversity and instability of educational data, the learning ability assessment model is easily affected by significant differences such as information gap, instability, and irrelevance between data, resulting in deviations in the model optimization results, thereby affecting the accuracy and reference value of the evaluation results.

[0023] In order to solve the problems in the related technology, the optimization method of the learning ability assessment model proposed in the present disclosure introduces a pre-step of data steady-state judgment to ensure the accuracy and reliability of the learning ability assessment model after constructing the student learning ability assessment model. After confirming that the data is in a steady state, the correlation between the prediction results of the two models is evaluated. If the preset conditions are met, the parameters of the learning ability assessment model are optimized to obtain a more accurate target model. It is possible to identify and eliminate unstable factors caused by data fluctuations or anomalies in advance, and to ensure that the parameters of the learning ability assessment model are optimized when the data is stable and the models work well together, effectively avoiding the optimization errors that may be caused by data instability, and greatly enhancing the accuracy and effectiveness of the evaluation model, providing more solid data support for the formulation of personalized education programs.

[0024] The following describes the optimization method, device, electronic device and storage medium of the learning ability assessment model of the embodiments of the present disclosure with reference to the accompanying drawings.

[0025] Figure 1 The following is a flow chart of an optimization method for a learning ability evaluation model provided by an embodiment of the present disclosure. Figure 1 As shown, the method includes: Step 101: Obtain a student's learning ability assessment model, which includes a learning achievement prediction model and a learning performance prediction model. The learning achievement prediction model is established based on multiple student achievement data sets, and the learning performance prediction model is established based on multiple student performance data sets.

[0026] In some embodiments, the learning ability assessment model consists of two core parts: a learning performance prediction model and a learning performance prediction model. The learning performance prediction model and the learning performance prediction model are trained and optimized based on different data sets to comprehensively assess the learning ability of students.

[0027] For the academic performance prediction model, the present disclosure may collect multiple student performance data sets. These student performance data sets may include factors that may affect the performance of multiple students, such as historical performance, test type, grade, class, gender, age, etc.

[0028] For the learning performance prediction model, the present disclosure can collect multiple student performance data sets. These student performance data sets can include multi-dimensional information such as students' class participation, homework completion, project performance, learning attitude, and teacher evaluation.

[0029] After collecting multiple student score data sets and multiple student performance data sets, the present invention can use the student score data sets and the student performance data sets to respectively train a learning score prediction model and a learning performance prediction model, and integrate the constructed learning score prediction model and the learning performance prediction model into a learning ability assessment model.

[0030] Among them, when constructing a student learning ability assessment model, the learning performance prediction model and the learning performance prediction model can be simply merged by direct addition, that is, the prediction results of the two models are directly added as the basis for comprehensive evaluation; the contribution of different aspects to students' learning ability can also be considered, and different weights can be assigned to the learning performance prediction model and the learning performance prediction model, and then weighted addition can be performed. The specific merging method can be based on actual needs and is not limited in the present disclosed embodiments.

[0031] After merging the learning achievement prediction model and the learning performance prediction model, a comprehensive learning ability assessment model can be obtained, which can comprehensively and accurately assess students' learning ability.

[0032] Step 102: Determine whether the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state respectively.

[0033] In some embodiments, in order to ensure that the data set used to train the learning ability assessment model is stable, that is, the data has no significant trend changes or periodic fluctuations, the present disclosure can determine whether multiple student score data sets and multiple student performance data sets are in a steady state based on the change trend of the student score and performance data over time. If the data set is determined to be in a steady state, step 103 can be continued; if not, the present disclosure needs to collect data again or adjust the data collection strategy.

[0034] It is understandable that the present disclosure can determine whether a data set is in a steady state by presetting a preset time interval and a preset threshold for each data set, that is, calculating the change rate of each data set within the preset time interval, and comparing it with the preset threshold to determine whether each data set is in a steady state. Subsequently, based on the number of data sets in a steady state, the overall steady state of multiple student achievement data sets and multiple student performance data sets is evaluated.

[0035] Step 103: If the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state, determine the correlation between the learning achievement prediction results output by the learning achievement prediction model and the learning performance prediction results output by the learning performance prediction model.

[0036] In some embodiments, after the present disclosure determines that multiple student performance data sets and multiple student performance data sets are in a steady state as a whole, the present disclosure can determine the correlation between the learning performance prediction model and the learning performance prediction model. The correlation between the learning performance prediction model and the learning performance prediction model represents the degree of association between the outputs of the learning performance prediction model and the learning performance prediction model, which helps to evaluate whether the two models consistently reflect the student's learning ability.

[0037] The present invention can input a stable learning performance prediction data set and a student performance data set into a learning performance prediction model and a learning performance prediction model respectively, thereby obtaining learning performance prediction results and learning performance prediction results output by the learning performance prediction model and the learning performance prediction model.

[0038] Afterwards, the present disclosure uses statistical methods (such as Pearson correlation coefficient and Spearman rank correlation coefficient) to calculate the correlation between the learning performance prediction results and the learning performance prediction results. If the correlation meets the preset conditions (such as the correlation reaches a certain preset correlation threshold), it indicates that the two models have good consistency in evaluating students' learning ability; if not, it is determined that the current learning performance prediction results and learning performance prediction results are abnormal, the results are invalidated, and the reasons are further analyzed, such as model design problems, improper feature selection, etc.

[0039] Step 104: If the correlation meets the preset conditions, the parameters of the learning ability evaluation model are optimized to obtain the target model.

[0040] In some embodiments, the present disclosure improves the prediction accuracy and generalization ability of the learning ability assessment model by adjusting model parameters.

[0041] The parameter tuning disclosed in the present invention can use methods such as gradient descent, cross validation, grid search, random search or Bayesian optimization to find the optimal model parameter combination.

[0042] The present invention obtains the final optimized learning ability assessment model (i.e., target model) by continuously evaluating the performance of the optimized model and further adjusting the parameters or model structure according to the evaluation results. The target model should have higher accuracy and reliability in predicting students' academic performance and results.

[0043] In summary, the technical solution provided by the present invention obtains a student's learning ability assessment model, the learning ability assessment model includes a learning performance prediction model and a learning performance prediction model, the learning performance prediction model is established based on multiple student performance data sets, and the learning performance prediction model is established based on multiple student performance data sets; respectively determine whether multiple student performance data sets and multiple student performance data sets are in a steady state; if multiple student performance data sets and multiple student performance data sets are in a steady state, determine the correlation between the learning performance prediction results output by the learning performance prediction model and the learning performance prediction results output by the learning performance prediction model; if the correlation meets the preset conditions, optimize the parameters of the learning ability assessment model to obtain the target model, and introduce the steady-state judgment of the data and the correlation judgment of the model output results before the parameter optimization process of the learning ability assessment model to ensure the stability and correlation of the data, avoid the problem of inaccurate parameter optimization due to data instability and low correlation between data, improve the accuracy and effectiveness of the learning ability assessment model, and provide schools and parents with more comprehensive and accurate student learning ability assessment services, thereby helping schools and parents to formulate and implement effective learning tutoring plans.

[0044] As a possible implementation, Figure 2 The flowchart of another optimization method of the learning ability evaluation model shown in the figure, based on the above embodiment, the specific process of determining whether multiple student score data sets and multiple student performance data sets are in a steady state respectively includes the following steps: Step 201: respectively obtain a preset time interval and a preset threshold value corresponding to each of a plurality of student score data sets and a plurality of student performance data sets.

[0045] In some embodiments, the present disclosure can pre-set a corresponding preset time interval and preset threshold for each data set (whether it is a student score data set or a student performance data set).

[0046] The preset time interval defines the time range for observing changes in the data set, such as a semester, a school year, or a shorter period of time.

[0047] The preset threshold defines the acceptable upper limit of the change rate of the data set within the time interval, that is, if the change rate exceeds this value, the data set is considered unstable.

[0048] Step 202: for each data set, if the change rate of each data set in the corresponding preset time interval is less than or equal to the corresponding preset threshold, it is determined that the data set is in a steady state.

[0049] In some embodiments, the present disclosure can calculate the rate of change of each data set within a preset time interval, which can be achieved by comparing the statistics (such as mean, median, etc.) at the beginning and end of the time interval, and calculating the relative change percentage.

[0050] If the rate of change of a data set within a preset time interval is less than or equal to the corresponding preset threshold, the data set is considered to be in a steady state within the time period.

[0051] Among them, for each data set, the preset thresholds of different data types contained therein can be the same or different, depending on the actual situation and is not limited in the embodiments of the present disclosure.

[0052] In an optional embodiment of the present disclosure, the preset threshold of the present disclosure can be set using different score levels of a data set, so as to calculate the evaluation index corresponding to each score level based on the actual data to reflect the data stability level, thereby more accurately grasping the data status. That is, according to different score levels, the data scores in each data set are divided into grades to obtain different score levels corresponding to different data in each data set; for each score level, the data scores in the score level and the average of all data scores in the score level are used to calculate the evaluation index corresponding to each data score in the score level; the evaluation indexes corresponding to all data scores in the score level are compared, and the evaluation index with the smallest evaluation index corresponding to the data score is used as the preset stability threshold (i.e., preset threshold) of the score level.

[0053] Specifically, for a student score data set or a student performance data set, the present disclosure can divide the data set into q levels according to different score levels, expressed as M=[M1,M2,...Mq], where M represents the data score divided into the corresponding score level.

[0054] Calculate the evaluation index J (Mn) of each M level respectively, the expression is:

[0055] Among them, Mi is the i-th data score in a certain score level data sample, is the mean of all data scores in a data sample of a certain score level, and n is the amount of data in a data sample of a certain level.

[0056] Take minJ (Mn) with the minimum evaluation index under this score level as the preset steady-state threshold (i.e., the preset threshold) of the score level corresponding to M.

[0057] Step 203: If the number of data sets in a steady state among the plurality of student score data sets and the plurality of student performance data sets meets a preset number condition, it is determined that the plurality of student score data sets and the plurality of student performance data sets are in a steady state.

[0058] In some embodiments, the preset quantity condition can be flexibly set according to the actual application scenario and data characteristics.

[0059] The preset quantity condition in the present disclosure may refer to the total number of multiple student score data sets and multiple student performance data sets, that is, all student score data sets and all student performance data sets must be in a steady state before the overall data set can be considered to be in a steady state.

[0060] The preset quantity condition in the present disclosure may also refer to a partial number or partial proportion of the overall data set of multiple student score data sets and multiple student performance data sets, that is, whether it is a student score data set or a student performance data set, when the number of steady-state data sets meets a certain number or the proportion of the number of steady-state data sets to the total number of data sets reaches a certain threshold, it is considered that the student score data set and the student performance data set are both in a steady state.

[0061] The preset quantity condition in the present disclosure can also be the partial quantity or partial proportion corresponding to each of the multiple student score data sets and the multiple student performance data sets, that is, when the number of student score data sets in a steady state meets a certain number or the proportion of the number of data sets in a steady state to the total number of data sets reaches a certain threshold, and the number of student performance data sets in a steady state meets a certain number or the proportion of the number of data sets in a steady state to the total number of data sets reaches a certain threshold, the student score data sets and the student performance data sets are considered to be in a steady state. This setting method is more sophisticated and can more accurately reflect the characteristics of different data sets.

[0062] In summary, the present invention can effectively narrow data differences and reduce errors caused by data instability to cope with complex and changeable educational data by adding data steady-state judgment before parameter optimization of the learning ability assessment model.

[0063] As a possible implementation, Figure 3 The flowchart of another optimization method of the learning ability evaluation model shown in the figure, based on the above embodiment, the specific process of determining the correlation between the learning performance prediction result output by the learning performance prediction model and the learning performance prediction result output by the learning performance prediction model includes the following steps: Step 301: Based on multiple student performance data sets, use the learning performance prediction model to determine multiple learning performance prediction results.

[0064] In some embodiments, the present disclosure may use a learning performance prediction model to predict multiple student performance data sets to obtain multiple learning performance prediction results.

[0065] Step 302: Based on multiple student performance data sets, use a learning performance prediction model to determine multiple learning performance prediction results.

[0066] In some embodiments, the present disclosure may use a learning performance prediction model to predict multiple student performance data sets to obtain multiple learning performance prediction results.

[0067] Step 303: respectively determine the score changes and statistical values ​​of the score changes for multiple learning score prediction results within a preset time interval, and the performance changes and statistical values ​​of the performance changes for multiple learning performance prediction results.

[0068] In some embodiments, for each student within a preset time period (such as one semester, one school year, etc.), the change in the academic performance prediction results and the change in the learning performance prediction results of each student are calculated.

[0069] Conduct statistical analysis on the changes in grades of all students, and calculate the statistical values ​​of grade changes and performance changes (such as mean, standard deviation, median, etc.) to understand the overall trend and distribution of grade changes.

[0070] Step 304: Determine the correlation based on the score change, the statistical value of the score change, the performance change, and the statistical value of the performance change.

[0071] In some embodiments, if there is a lack of sufficient correlation between the data sets, the learning performance prediction model and the learning performance prediction model may encounter difficulties in the fusion process, resulting in poor fitting of the overall model, which ultimately affects the prediction performance. Therefore, before the parameter optimization process of the learning ability evaluation model, the present disclosure can use correlation calculation to evaluate the correlation between the learning performance prediction results and the learning performance prediction results to determine whether the current model is suitable for fusion training, thereby improving the accuracy and robustness of the prediction.

[0072] Specifically, the present disclosure may determine the correlation according to the following formula 1: Formula 1 in, Indicates relevance, Indicates the change in performance. It represents the average value of the change in grades. Indicates the amount of change in performance, represents the average value of the performance variation, Indicates the amount of data.

[0073] In summary, the present invention determines the correlation between the learning performance prediction results and the learning grade prediction results before optimizing the parameters of the learning ability assessment model to ensure the synergy between the learning performance prediction model and the learning grade prediction model, thereby improving the effect of model fusion and helping schools and parents to more accurately understand the all-round development of students' learning ability.

[0074] As a possible implementation, Figure 3 The flowchart of another optimization method of the learning ability evaluation model shown in the figure, based on the above embodiment, if the correlation meets the preset conditions, the parameters of the learning ability evaluation model are optimized to obtain the specific process of the target model, including the following steps: Step 401: If the correlation is greater than zero, it is determined that the learning performance prediction result output by the learning performance prediction model is positively correlated with the learning performance prediction result output by the learning performance prediction model.

[0075] In some embodiments, if >0, indicating that the learning performance prediction results output by the learning performance prediction model and the learning performance prediction results output by the learning performance prediction model are positively correlated.

[0076] Step 402: multiple student achievement data sets and multiple student performance data sets are used as data samples, and data labels corresponding to the data samples are obtained, where the data labels are actual learning ability scores of each student in the data samples.

[0077] In some embodiments, the above-mentioned multiple stable student achievement data sets and multiple student performance data sets are used as data samples, and data labels corresponding to the data samples are obtained through external evaluation (such as teacher evaluation, standardized test, etc.), that is, the actual learning ability score of each student.

[0078] Step 403: Establish a regression model, where the regression model includes a parameter matrix of the learning ability evaluation model.

[0079] In some embodiments, the present disclosure may use data samples and data labels to construct a regression model, which is intended to predict students' learning abilities by learning features in the data samples.

[0080] Among them, in the present disclosure, the regression model may include a parameter matrix of the learning ability assessment model. The parameter matrix is ​​a key part of the regression model that needs to be optimized during the training process, and they determine how the regression model learns the student's learning ability score from the input data (i.e., the student score data set and the student performance data set).

[0081] Specifically, the regression model established in the present disclosure is as follows: Formula 2 in, represents the data label, , represents the actual learning ability score of the nth student, represents a data sample, represents the parameter matrix, and the elements in the parameter matrix are the partial regression coefficients in the learning achievement prediction model and the learning performance prediction model. Indicates the error value.

[0082] In an optional embodiment, the learning performance prediction model may be:

[0083] in, Indicates the evaluation results of students’ academic performance. represents a constant term, represents the historical test scores in the student score dataset, represents the daily homework scores in the student score dataset, represents the teacher rating in the student performance dataset, represents the classroom performance score in the student performance dataset, Represents the partial regression coefficient of the learning performance prediction model.

[0084] Learning performance prediction models can be:

[0085] in, Indicates the evaluation results of students’ learning performance, represents a constant term, represents the learning environment in the student performance dataset, represents the learning state in the student performance dataset, represents the learning style in the student performance dataset, represents the learning behavior in the student performance dataset, represents the partial regression coefficient of the learning performance prediction model.

[0086] The present disclosure can directly add the learning achievement prediction model Y and the learning performance prediction model Z to construct a regression model for evaluating students' learning ability.

[0087] The following formula can be used to construct the regression model:

[0088] in represents the evaluation results of students’ learning ability of the regression model, represents the evaluation result of the student's academic performance of the academic performance prediction model, Represents the evaluation result of the student learning performance of the learning performance prediction model.

[0089] Optionally, the present disclosure may also introduce weights in the process of constructing a regression model, that is, assigning different weights to the learning performance prediction model and the learning performance prediction model to construct a regression model.

[0090] Among them, the following formula can be used to introduce weights to construct a regression model:

[0091] in represents the evaluation results of students’ learning ability of the regression model, represents the evaluation result of the student's academic performance of the academic performance prediction model, Represents the evaluation result of the student learning performance of the learning performance prediction model. and is the weight coefficient. 1, or and Any preset value.

[0092] The present disclosure can set different weight values ​​for learning results and learning performance according to the focus of student ability assessment. For example, if the student's performance focuses on objective aspects, the weight value can be set. > ; Focus on the subjective aspect, you can set < ; When evaluating objective and subjective aspects in a balanced way, you can set = The above are not limited in the embodiments of the present disclosure.

[0093] Therefore, the partial regression coefficient in the parameter matrix in the embodiment of the present disclosure may specifically refer to , and learning performance prediction models .

[0094] Step 404: Use the data samples and data labels to train the regression model until the training is completed, and obtain the optimized parameter matrix of the learning ability evaluation model.

[0095] In some embodiments, the present disclosure may use the above-determined data samples and data labels to train the regression model. The training process may include multiple iterations, and each iteration adjusts the parameters of the model according to the difference between the predicted learning ability score value of the current regression model and the actual learning ability score value in the data label. The difference between the predicted learning ability score value of the regression model and the actual learning ability score value in the data label is quantified in the present disclosure by a loss function.

[0096] The present disclosure can use the gradient descent method to train the regression model to adjust its parameters. This process will be repeated until the loss value meets the preset loss condition (that is, the loss value is reduced to below the preset loss value or the number of training times reaches the preset number of iterations), at which time the regression model is considered to have been trained. After the training is completed, the present disclosure can obtain the optimized parameter matrix of the learning ability assessment model. The optimized parameter matrix contains the optimal parameter values ​​(that is, the optimal partial regression coefficients) learned by the regression model during the training process. The optimized parameter matrix defines how the learning ability assessment model predicts the student's learning ability score from the input features. The optimal partial regression coefficients in the optimized parameter matrix can make the learning ability assessment model have higher prediction accuracy.

[0097] The loss function established in the present disclosure is as follows: Formula 3 in, represents the loss value, represents the actual learning ability score of the mth student, represents the predicted learning ability score of the mth student calculated using the learning ability assessment model; The regression model is trained using data samples and data labels until the loss value meets the preset loss condition. The training ends and the optimized parameter matrix of the learning ability evaluation model is obtained.

[0098] Use data samples and data labels to train the regression model until the loss value meets the preset loss condition. The training ends and the optimization parameter matrix of the learning ability evaluation model is obtained, including: During the training process, the parameter matrix of the learning ability evaluation model is expressed as the following formula 4: Formula 4; The optimization parameter matrix is ​​expressed as the following formula 5: Formula 5 in represents the learning rate, Represents the partial derivative of the loss value G; Using the gradient descent method, when the loss value meets the preset loss condition, the parameter matrix corresponding to the loss value that meets the preset loss condition is determined as the optimized parameter matrix.

[0099] Step 405: Use the optimized parameter matrix to update the learning ability evaluation model to obtain a target model.

[0100] In some embodiments, the present disclosure may use the optimized parameter matrix obtained through training to update the learning ability evaluation model. The updated learning ability evaluation model (ie, the target model) can better evaluate the learning ability of students.

[0101] In summary, the present invention further improves the accuracy of the evaluation results of the learning ability assessment model by optimizing the parameters of the learning ability assessment model while ensuring that the learning score prediction results and the learning performance prediction results are relevant, providing schools and parents with more comprehensive and accurate student learning ability assessment services, thereby helping schools and parents to formulate and implement effective learning tutoring programs.

[0102] Corresponding to the above-mentioned optimization method of the learning ability evaluation model, the present invention also proposes an optimization device for the learning ability evaluation model. Since the device embodiment of the present invention corresponds to the above-mentioned method embodiment, the details not disclosed in the device embodiment can be referred to the above-mentioned method embodiment, and will not be repeated in the present invention.

[0103] Figure 5 A schematic diagram of a structure of an optimization device for a learning ability evaluation model provided by an embodiment of the present disclosure, such as Figure 5 As shown, the device comprises: An acquisition unit 510 is used to acquire a student's learning ability assessment model, wherein the learning ability assessment model includes a learning performance prediction model and a learning performance prediction model, wherein the learning performance prediction model is established based on multiple student performance data sets, and the learning performance prediction model is established based on multiple student performance data sets; a steady state determination unit 520, for respectively determining whether the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state; A correlation determination unit 530, for determining the correlation between the learning performance prediction results output by the learning performance prediction model and the learning performance prediction results output by the learning performance prediction model if the plurality of student performance data sets and the plurality of student performance data sets are in a steady state; The optimization unit 540 is used to optimize the parameters of the learning ability evaluation model to obtain a target model if the correlation meets the preset conditions.

[0104] In some embodiments of the present disclosure, the steady state determination unit 520 is configured to: Respectively obtain a preset time interval and a preset threshold value corresponding to each of the multiple student score data sets and the multiple student performance data sets; For each data set, if the rate of change of each data set in the corresponding preset time interval is less than or equal to the corresponding preset threshold, it is determined that the data set is in a steady state; If the number of data sets in a steady state among the plurality of student achievement data sets and the plurality of student performance data sets meets a preset number condition, it is determined that the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state.

[0105] In some embodiments of the present disclosure, the correlation determination unit 530 is configured to: Based on multiple student performance data sets, using the learning performance prediction model, determining multiple learning performance prediction results; Based on multiple student performance data sets, using a learning performance prediction model, determining multiple learning performance prediction results; Respectively determine the score changes and statistical values ​​of the score changes for a plurality of learning score prediction results within a preset time interval, and the performance changes and statistical values ​​of the performance changes for a plurality of learning performance prediction results; Correlation is determined based on the change in grades, the statistical value of the change in grades, the change in performance, and the statistical value of the change in performance.

[0106] In some embodiments of the present disclosure, the correlation determination unit 530 is configured to: The correlation is determined according to the following formula 1: Formula 1 in, Indicates relevance, Indicates the change in performance. It represents the average value of the change in grades. Indicates the amount of change in performance, represents the average value of the performance variation, Indicates the amount of data.

[0107] In some embodiments of the present disclosure, the optimization unit 540 is used to: If the correlation is greater than zero, it is determined that the learning performance prediction result output by the learning performance prediction model is positively correlated with the learning performance prediction result output by the learning performance prediction model; Taking multiple student score data sets and multiple student performance data sets as data samples, and obtaining data labels corresponding to the data samples, the data labels are actual learning ability scores of each student in the data samples; Establishing a regression model, the regression model includes a parameter matrix of a learning ability assessment model; Use data samples and data labels to train the regression model until the training is completed, and obtain the optimized parameter matrix of the learning ability evaluation model; The learning ability evaluation model is updated using the optimized parameter matrix to obtain the target model.

[0108] In some embodiments of the present disclosure, the optimization unit 540 is used to: The regression model is established as shown in the following formula 2: Formula 2 in, represents the data label, , represents the actual learning ability score of the nth student, represents a data sample, represents the parameter matrix, and the elements in the parameter matrix are the partial regression coefficients in the learning achievement prediction model and the learning performance prediction model. Indicates the error value.

[0109] In some embodiments of the present disclosure, the optimization unit 540 is used to: The loss function is established as follows: Formula 3 in, represents the loss value, represents the actual learning ability score of the mth student, represents the predicted learning ability score of the mth student calculated using the learning ability assessment model; The regression model is trained using data samples and data labels until the loss value meets the preset loss condition. The training ends and the optimized parameter matrix of the learning ability evaluation model is obtained.

[0110] In some embodiments of the present disclosure, the optimization unit 540 is used to: During the training process, the parameter matrix of the learning ability evaluation model is expressed as the following formula 4: Formula 4; The optimization parameter matrix is ​​expressed as the following formula 5: Formula 5 in represents the learning rate, Represents the partial derivative of the loss value G; Using the gradient descent method, when the loss value meets the preset loss condition, the parameter matrix corresponding to the loss value that meets the preset loss condition is determined as the optimized parameter matrix.

[0111] It should be noted that the above explanation of the method embodiment is also applicable to the device of this embodiment, and the principle is the same, which is not limited in this embodiment.

[0112] Based on the above Figure 1 , Figure 2 , Figure 3 as well as Figure 4The method shown in the figure, accordingly, this embodiment also provides a computer program product, including a computer program, which implements the above-mentioned Figure 1 , Figure 2 , Figure 3 as well as Figure 4 The method shown.

[0113] Based on the above Figure 1 , Figure 2 , Figure 3 as well as Figure 4 The method shown in the embodiment also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned Figure 1 , Figure 2 , Figure 3 as well as Figure 4 The method shown.

[0114] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, USB flash drive, mobile hard disk, etc.), including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of the present application.

[0115] like Figure 6 The figure shows a hardware structure diagram of an electronic device of the present invention, including: at least one processor 601; and, A memory 602 is communicatively connected to at least one of the processors 601; wherein, The memory 602 stores instructions that can be executed by at least one of the processors, and the instructions are executed by at least one of the processors to enable at least one of the processors to perform the optimization method of the learning ability assessment model as described above.

[0116] Figure 6 A processor 601 is taken as an example.

[0117] The electronic device may further include: an input device 603 and a display device 604 .

[0118] The processor 601, the memory 602, the input device 603 and the display device 604 may be connected via a bus or other means, and the figure takes the connection via a bus as an example.

[0119] The memory 602 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the review content generation method in the embodiment of the present application, for example, Figure 1 , Figure 2 , Figure 3 as well as Figure 4 The processor 601 executes various functional applications and data processing by running the non-volatile software programs, instructions and modules stored in the memory 602, that is, the optimization method of the learning ability evaluation model in the above embodiment is implemented.

[0120] The memory 602 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required by at least one function; the data storage area may store data created according to the use of the review content generation method, etc. In addition, the memory 602 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some embodiments, the memory 602 may optionally include a memory remotely arranged relative to the processor 601, and these remote memories may be connected to the device for executing the review content generation method via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0121] The input device 603 can receive user clicks and generate signal inputs related to user settings and function controls of the review content generation method. The display device 604 can include display devices such as a display screen.

[0122] The one or more modules are stored in the memory 602, and when executed by the one or more processors 601, the optimization method of the learning ability evaluation model in any of the above method embodiments is executed.

[0123] Optionally, the above-mentioned physical device may also include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), etc.

[0124] Those skilled in the art will appreciate that the above-mentioned physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different arrangements of components.

[0125] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the above-mentioned physical device, and supports the operation of the information processing program and other software and / or programs. The network communication module is used to realize the communication between the components inside the storage medium, and the communication with other hardware and software in the information processing physical device.

[0126] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform, or by hardware. By applying the solution of this embodiment, compared with the current existing technology, this embodiment obtains a student's learning ability assessment model, and the learning ability assessment model includes a learning performance prediction model and a learning performance prediction model. The learning performance prediction model is established based on multiple student performance data sets, and the learning performance prediction model is established based on multiple student performance data sets; respectively determine whether multiple student performance data sets and multiple student performance data sets are in a steady state; if multiple student performance data sets and multiple student performance data sets are in a steady state, determine the correlation between the learning performance prediction results output by the learning performance prediction model and the learning performance prediction results output by the learning performance prediction model; if the correlation meets the preset conditions, optimize the parameters of the learning ability assessment model to obtain the target model, and realize the introduction of data steady-state judgment and model output result correlation judgment before the parameter optimization process of the learning ability assessment model to ensure the stability and correlation of the data, avoid the problem of inaccurate parameter optimization due to data instability and low correlation between data, improve the accuracy and effectiveness of the learning ability assessment model, and provide schools and parents with more comprehensive and accurate student learning ability assessment services, thereby helping schools and parents to formulate and implement effective learning counseling plans.

[0127] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0128] The above is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features applied for herein.

Claims

1. An optimization method for a learning ability evaluation model, It is characterized in that include: Acquire a learning ability assessment model of the student, wherein the learning ability assessment model includes a learning performance prediction model and a learning performance prediction model, wherein the learning performance prediction model is established based on multiple student performance data sets, and the learning performance prediction model is established based on multiple student performance data sets; respectively determining whether the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state; If the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state, determining a correlation between the learning achievement prediction results output by the learning achievement prediction model and the learning performance prediction results output by the learning performance prediction model; If the correlation meets the preset conditions, the parameters of the learning ability evaluation model are optimized to obtain a target model.

2. The method according to claim 1, It is characterized in that The determining whether the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state comprises: Respectively obtain the preset time interval and the preset threshold value corresponding to each of the plurality of student score data sets and the plurality of student performance data sets; For each data set, if the change rate of each data set in the corresponding preset time interval is less than or equal to the corresponding preset threshold, it is determined that the data set is in a steady state; If the number of data sets in a steady state among the plurality of student achievement data sets and the plurality of student performance data sets meets a preset number condition, it is determined that the plurality of student achievement data sets and the plurality of student performance data sets are in a steady state.

3. The method according to claim 1, It is characterized in that Determining the correlation between the learning performance prediction result output by the learning performance prediction model and the learning performance prediction result output by the learning performance prediction model includes: Based on the multiple student performance data sets, using the learning performance prediction model, determine multiple learning performance prediction results; Based on the plurality of student performance data sets, using the learning performance prediction model, determining a plurality of learning performance prediction results; Respectively determine the score changes of the plurality of learning score prediction results within a preset time interval, the statistical value of the score changes, the performance changes of the plurality of learning performance prediction results, and the statistical value of the performance changes; The correlation is determined based on the score change amount, the statistical value of the score change amount, the performance change amount, and the statistical value of the performance change amount.

4. The method according to claim 3, It is characterized in that The determining of the correlation based on the score change, the statistical value of the score change, the performance change, and the statistical value of the performance change includes: The correlation is determined according to the following formula 1: Formula 1 in, represents the correlation, represents the change in the score, represents the average value of the change in the score, represents the amount of change in the performance, represents the average value of the performance variation, Indicates the amount of data.

5. The method according to any one of claims 1 to 4, It is characterized in that If the correlation satisfies a preset condition, the parameters of the learning ability evaluation model are optimized to obtain a target model, including: If the correlation is greater than zero, it is determined that the learning performance prediction result output by the learning performance prediction model is positively correlated with the learning performance prediction result output by the learning performance prediction model; Taking the multiple student achievement data sets and the multiple student performance data sets as data samples, and obtaining data labels corresponding to the data samples, wherein the data labels are actual learning ability scores of each student in the data samples; Establishing a regression model, wherein the regression model includes a parameter matrix of the learning ability assessment model; Using the data samples and the data labels, training the regression model until the training is completed, and obtaining an optimized parameter matrix of the learning ability evaluation model; The learning ability evaluation model is updated using the optimization parameter matrix to obtain the target model.

6. The method according to claim 5, It is characterized in that The said establishing regression model comprises: The regression model is established as the following formula 2: Formula 2 in, represents the data label, , represents the actual learning ability score of the nth student, represents the data sample, represents the parameter matrix, the elements in the parameter matrix are the partial regression coefficients in the learning performance prediction model and the learning performance prediction model, Indicates the error value.

7. The method according to claim 5, It is characterized in that The use of the data samples and the data labels to train the regression model until the training is completed, and obtaining the optimization parameter matrix of the learning ability evaluation model includes: The loss function is established as follows: Formula 3 in, represents the loss value, represents the actual learning ability score of the mth student, represents the predicted learning ability score of the mth student calculated using the learning ability assessment model; The regression model is trained using the data samples and the data labels until the loss value meets the preset loss condition, then the training ends and an optimized parameter matrix of the learning ability evaluation model is obtained.

8. The method according to claim 7, It is characterized in that The regression model is trained using the data sample and the data label until the loss value meets the preset loss condition, and the training is terminated. The optimization parameter matrix of the learning ability evaluation model is obtained, including: During the training process, the parameter matrix of the learning ability evaluation model is expressed as the following formula 4: Formula 4: The optimization parameter matrix is ​​expressed as the following formula 5: Formula 5 in represents the learning rate, Represents the partial derivative of the loss value G; By using the gradient descent method, when the loss value satisfies the preset loss condition, the parameter matrix corresponding to the loss value that satisfies the preset loss condition is determined as the optimization parameter matrix.

9. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

10. An electronic device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, It is characterized in that When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.