Parental-end applet system for recommending and reporting learning conditions of students based on parent portraits

Abstract

The invention provides a parent terminal applet system for recommending and reporting student learning conditions based on parent portraits. The system comprises a data acquisition module for acquiring a student multi-source learning data set from students after a parent terminal applet logs in; meanwhile, parent portrait information is recommended and determined for parents of the students; the fusion processing module performs data processing on the multi-source learning data set of the student by adopting a data fusion method; the behavior analysis module performs learning behavior analysis according to the student multi-source learning processing data to obtain student learning analysis data; the report generation module is used for generating a learning condition report text based on student learning analysis data in combination with parent portrait information through a natural language processing model to obtain learning condition report text information; and the condition feedback module feeds back the learning condition of the student in the parent-side applet according to the learning condition report text information. According to the invention, learning conditions of students can be reflected more deeply, and parents can be ensured to accurately understand learning condition feedback information of the students.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of educational informatization, in particular to a parent terminal applet system for recommending and reporting student learning conditions based on parent portraits. BACKGROUND

[0002] In today's educational environment, parents pay more and more attention to student learning conditions. Traditional methods of home-school communication, such as parent-teacher meetings and phone communication, have problems such as information lag and incomplete communication. With the rapid development of Internet technology and the widespread popularity of mobile devices, more and more parents begin to use various online education tools and platforms to monitor and assist students' learning. However, existing parent terminal applets only provide scattered learning data, making it difficult for parents to quickly gain meaningful insights from them. Moreover, the learning data is only simply summarized, which cannot deeply analyze students' learning behavior and reveal the underlying problems of their knowledge mastery. Therefore, the present application proposes a parent terminal applet system for recommending and reporting student learning conditions based on parent portraits to solve the problems raised in the background. SUMMARY

[0003] The present application aims to provide a parent terminal applet system for recommending and reporting student learning conditions based on parent portraits, which not only realizes the arrangement of multi-source data and more deeply reflects students' learning conditions, but also combines parent portraits for feedback consideration, enabling parents to efficiently understand students' learning conditions, thereby effectively improving the quality and effectiveness of family education and realizing the good connection and cooperation between family education and school education, providing stronger support and protection for the overall development of students.

[0004] To achieve the above-mentioned purpose, the present application provides the following technical solution: a parent terminal applet system for recommending and reporting student learning conditions based on parent portraits, comprising:

[0005] A data acquisition module is used to acquire multi-source learning data for students after logging into the parent terminal applet, obtaining a student multi-source learning data set; and recommend a parent portrait for the student's parents, determining parent portrait information;

[0006] A fusion processing module is used to process data for the student multi-source learning data set using a data fusion method, obtaining student multi-source learning processing data;

[0007] A behavior analysis module is used to analyze learning behavior based on the student multi-source learning processing data, obtaining student learning analysis data;

[0008] A report generation module is used to generate a learning condition report text based on the student learning analysis data and the parent portrait information using a natural language processing model, obtaining learning condition report text information;

[0009] A case feedback module is configured to report learning case text information in the parent terminal applet.

[0010] Preferably, the data acquisition module acquires multi-source learning data of the student, including:

[0011] Determining a target student, and collecting associated information based on the target student on the online education platform to obtain an original data set;

[0012] Performing integrity analysis and processing on the original data set to obtain a student multi-source learning data set.

[0013] Preferably, the fusion processing module adopts a data fusion method to process the student multi-source learning data set, including:

[0014] Performing data cleaning on the student multi-source learning data set to obtain a first processed data set;

[0015] Analyzing the data format of the first processed data set and performing standardized processing on the data format to obtain a second processed data set;

[0016] Performing fusion processing in the second processed data set to determine integrated data information;

[0017] Performing standardized processing on the integrated data information to obtain student multi-source learning processing data.

[0018] Preferably, the behavior analysis module analyzes learning behavior according to the student multi-source learning processing data, including:

[0019] A behavior analysis unit: based on the student multi-source learning processing data, analyzes and calculates the knowledge point mastery rate of the student to obtain first learning analysis data; analyzes the student multi-source learning processing data to determine the learning duration and homework submission frequency of the student, and performs learning attitude clustering analysis according to the learning duration and homework submission frequency to determine the learning attitude of the student to obtain second analysis data; combines the first learning analysis data with the learning duration to predict the learning behavior of the student to obtain third analysis data;

[0020] The analysis diagnosis unit is configured to perform learning diagnosis on the student based on the first learning analysis data, including: judging the first learning analysis data in combination with a first threshold to obtain a first judgment result; obtaining error answering data based on the first judgment result, and performing scale judgment on the error answering data in combination with a second threshold to obtain a second judgment result; expanding the error answering data according to the second judgment result to obtain perfect data information; performing error type analysis on the perfect data information by using a decision tree algorithm model to determine error type distribution information; analyzing the relevance between the error type and the knowledge point based on the error type distribution information, and determining associated knowledge points to obtain a weak knowledge point set; and generating a family education guidance suggestion based on the weak knowledge point set.

[0021] Preferably, the report generation module generates a learning situation report text based on the student learning analysis data in combination with the parent portrait information by using a natural language processing model, including:

[0022] Performing structured template matching on the student learning analysis data to determine a structured target template;

[0023] Generating an initial learning situation report text information based on the student learning analysis data by using a natural language processing model based on the structured target template;

[0024] Adjusting the style of the initial learning situation report text information in combination with the parent portrait information to obtain adjusted learning situation report text information;

[0025] Optimizing the adjusted learning situation report text information to obtain final learning situation report text information.

[0026] Preferably, adjusting the style of the initial learning situation report text information in combination with the parent portrait information includes:

[0027] Determining target parent portrait information by using a parent portrait recommendation;

[0028] Performing cognitive level analysis on the parent based on the target parent portrait information to obtain a first feature of the parent;

[0029] Performing personality habit analysis on the parent based on the target parent portrait information to obtain a second feature of the parent;

[0030] Adjusting the expression of the initial learning situation report text information according to the second feature of the parent to obtain first adjusted learning situation report text information;

[0031] Enriching the semantics of the first adjusted learning situation report text information according to the first feature of the parent to obtain second adjusted learning situation report text information.

[0032] Preferably, when optimizing the learning condition adjustment report text information, the learning condition adjustment report text information is analyzed and optimized for sentence coherence to obtain learning condition adjustment report text first optimization information; the completeness of the learning condition adjustment report text first optimization information is analyzed and judged, and when the learning condition adjustment report text first optimization information is complete, the learning condition adjustment report text first optimization information is sequentially analyzed for ambiguity description, the content with ambiguity description is locked, the content with ambiguity description is corrected through semantic analysis, the learning condition adjustment report text second optimization information is determined, and the final learning condition report text information is obtained.

[0033] Preferably, before the report generation module generates the learning condition report text based on the student learning analysis data and in combination with the parent portrait information through the natural language processing model, the student learning analysis data is analyzed and set for optimization, including:

[0034] Identifying and analyzing the weaknesses of the student to determine the type of the weaknesses of the student;

[0035] Starting the psychological sensitivity evaluation engine based on the type of the weaknesses of the student to obtain the psychological sensitivity of the student to the type of the weaknesses of the student and determine the psychological sensitivity analysis result;

[0036] Optimizing the student learning analysis data according to the psychological sensitivity analysis result, creating a hidden space, and determining the student learning analysis optimization data;

[0037] Determining the self-improvement probability through the self-improvement possibility prediction engine for the hidden space;

[0038] According to the self-improvement probability and the psychological sensitivity to the type of the weaknesses of the student, the hidden time of the hidden space is determined through the hidden time calculation algorithm, and the hidden space is set according to the hidden time;

[0039] Generating the learning condition report text based on the hidden time; when the hidden time is valid, the learning condition report text is generated based on the student learning analysis optimization data and in combination with the parent portrait information through the natural language processing model, and the learning condition report text information is annotated according to the hidden space; at the same time, the hidden space is supervised, and an emergency report is made according to the supervision result; when the hidden time is invalid, the hidden space is unhidden, and the learning condition report text is generated based on the hidden space and in combination with the parent portrait information through the natural language processing model.

[0040] Preferably, when the situation feedback module feeds back the student learning condition according to the learning condition report text information, the historical feedback information is updated for feedback, including:

[0041] The learning condition report text information is segmented and disassembled to obtain a plurality of information blocks;

[0042] According to the information block, the integrity analysis is combined with the historical feedback information, the corresponding relationship is determined, the missing information is determined, and the integrity analysis result is obtained;

[0043] According to the integrity evaluation result, the corresponding information block is marked according to the corresponding relationship and the historical feedback information, and the corresponding historical feedback information is obtained for the missing information, and the missing information is supplemented according to the corresponding historical feedback information, and the missing information block is obtained.

[0044] Integrate the marked information block and the missing information block, and perform personalized reporting based on the integration result.

[0045] Preferably, the situation feedback module also adjusts the learning situation report text information in real time, including:

[0046] Identify and analyze the reporting object to determine whether the reporting object is consistent with the current parent portrait, and obtain an identification analysis result;

[0047] When the identification analysis result is that the reporting object is inconsistent with the current parent portrait, a feedback adjustment request instruction is generated;

[0048] Based on the feedback adjustment request, the learning situation report text information is adjusted by the report generation module, and the adjusted learning situation report text information is obtained;

[0049] The adjusted learning situation report text information is fed back to the student learning situation in the situation feedback module.

[0050] The present application realizes efficient comprehensive data arrangement of student learning situation, reflects student's deep learning situation, improves the comprehensiveness of student learning situation report, and combines parent portrait to determine learning situation report text information recommendation, enhances the readability and understanding of the report, so that parents can efficiently understand the learning situation of students, thereby effectively improving the quality and effect of family education, realizing good connection and cooperation of family education and school education, and providing more powerful support and protection for the comprehensive development of students.

[0051] Other features and advantages of the present application will be described in the following specification, and some will become apparent from the specification, or will be understood by those skilled in the art. The purpose and other advantages of the present application can be achieved and obtained by the structure specifically pointed out in the application file.

[0052] The technical solutions of the present application will be further described in detail below with the help of the drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0053] The accompanying drawings are included to provide a further understanding of the application, and are incorporated in and constitute a part of this specification, illustrate embodiments of the application, and together with the description serve to explain the application, and, to specify the application, in which:

[0054] Figure 1 A schematic diagram of the parent terminal applet system according to the present application;

[0055] Figure 2 A schematic diagram of the behavior analysis module in the parent terminal applet system according to the present application;

[0056] Figure 3 A schematic diagram of the behavior analysis module in the parent terminal applet system according to the present application;

[0057] Figure 4 A schematic diagram of the report generation module in the parent terminal applet system according to the present application. DETAILED DESCRIPTION

[0058] The preferred embodiments of the present application will be described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to explain and illustrate the present application, and are not used to limit the present application.

[0059] As shown in Figure 1 , the parent terminal applet system for recommending and reporting student learning conditions based on parent portraits according to the embodiments of the present application comprises:

[0060] A data acquisition module, configured to acquire multi-source learning data of a student to obtain a student multi-source learning data set after a parent terminal applet is logged in, and to recommend a parent portrait for a parent of the student to determine parent portrait information.

[0061] The data acquisition module needs to determine a target student after the parent terminal applet is logged in by the parent, so as to acquire multi-source learning data of the target student. Here, the data acquisition module comprises a first acquisition unit and a second acquisition unit. The first acquisition unit acquires multi-source learning data of the student to obtain a student multi-source learning data set, and the second acquisition unit performs recommendation based on the parent portrait to determine an object of feedback of the student learning condition, so as to determine a target reporting object in the parent portrait and obtain parent portrait information. The multi-source learning data set comprises learning duration, exercise completion condition, test score, etc. The parent portrait information comprises self-education level, education philosophy, expectation for the student, participation degree in the student learning process, learning direction of interest, personality habit, etc.

[0062] A fusion processing module, configured to perform data processing on the student multi-source learning data set by using a data fusion method to obtain student multi-source learning processing data.

[0063] The data processing is heterogeneous data processing.

[0064] The behavior analysis module is configured to analyze learning behaviors of the student according to the multi-source learning processing data of the student, and obtain learning analysis data of the student.

[0065] The behavior analysis module is configured to analyze learning behaviors of the student according to the multi-source learning processing data of the student, and obtain learning analysis data of the student.

[0066] The report generation module is configured to generate a learning condition report text based on the learning analysis data of the student and the parent portrait information by using a natural language processing model, and obtain learning condition report text information.

[0067] The report generation module is configured to generate a learning condition report text based on the learning analysis data of the student and the parent portrait information by using a natural language processing model, and obtain learning condition report text information.

[0068] The condition feedback module is configured to feed back the learning condition of the student according to the learning condition report text information in the parent terminal applet.

[0069] The condition feedback module is configured to feed back the learning condition of the student according to the learning condition report text information in the parent terminal applet.

[0070] The technical scheme realizes efficient comprehensive data arrangement of the learning situation of the student, reflects the deep learning situation of the student, improves the comprehensiveness of the learning situation report of the student, and further combines the parent portrait to determine the learning situation report text information recommendation, enhances the readability and understanding of the report, and enables the parent to efficiently understand the learning situation of the student, thereby effectively improving the quality and effect of family education, realizing good connection and cooperation of family education and school education, and providing stronger support and guarantee for the all-round development of the student. The data acquisition module obtains the learning data of the student from multiple sources, provides data support for the behavior analysis module, more accurately determines the learning situation of the student, and obtains the parent portrait information, so that the learning situation report text information can be determined in combination with the parent situation in the report generation module, the readability and understanding of the report are enhanced, the parent can efficiently and accurately understand the learning situation of the student, and the family education can be better, the effect of school-family cooperative education is improved, and the learning of the student is guaranteed. The fusion processing module processes the multi-source learning data set of the student, can eliminate the contradiction and redundancy between the data, obtains more accurate multi-source learning processing data, guarantees that the behavior analysis module can analyze the learning behavior according to the multi-source learning processing data of the student, and the behavior analysis module realizes multi-source data arrangement, comprehensively reflects the learning situation of the student from multiple data information, reflects the deep learning situation of the student, avoids the deviation of the learning situation of the student caused by a single factor, and also avoids the error of the analysis of the learning situation of the student caused by the appearance data, so that the learning behavior of the student can be accurately analyzed, the learning situation of the student can be more deeply reflected, the problems in the learning of the student can be found in time, the targeted learning improvement of the student is guaranteed, the effectiveness of family education is improved, and good connection and cooperation of family education and school education are realized, and stronger support and guarantee are provided for the all-round development of the student.

[0071] In one embodiment of the present application, the data acquisition module obtains multi-source learning data for the student, including:

[0072] The target student is determined, and the associated information of the target student on the online education platform is collected to obtain an original data set.

[0073] Wherein, based on the target student in online education platform for information collection, for students to watch video learning time, exercise completion and test results and other information acquisition, through embedded JavaScript script in online education platform real-time collection of students to watch video education time, get the first acquisition data information, and the first acquisition data information in JSON format storage to cloud database MongoDB, then based on the target student tracking exercise completion, in the back-end server to determine the student's exercise submission, and get the completion of the exercise results, then according to the completion of the exercise results of daily data integration, determine the exercise completion, get the second acquisition data information, at the same time, the second acquisition data information stored in the cloud database MongoDB corresponding storage unit, then based on the target student from the test system to get the score record, and using Pearson correlation coefficient algorithm analysis score and watch time, if the analysis results for positive correlation, according to the score record get the third acquisition data information, and the third acquisition data information stored in the cloud database MongoDB corresponding storage unit, otherwise, the third acquisition data information is blank. Here, the storage unit in the cloud database MongoDB is set according to the student, the corresponding storage unit refers to the corresponding target student storage unit.

[0074] The student multi-source learning dataset is obtained by performing integrity analysis and processing on the original dataset.

[0075] Wherein, when the integrity analysis and processing of the original dataset is performed, the integrity analysis of the original dataset is performed based on the storage unit to determine whether there is missing data in the original dataset, if there is missing data, the mean method is used to complete the missing data, so as to determine the student multi-source learning dataset based on the completed original dataset. Here, the student multi-source learning dataset is the student multi-source learning dataset of the target student.

[0076] The above data acquisition module can obtain more comprehensive learning information for students, so that the learning behavior analysis can more deeply reflect the learning situation of the students, and improve the depth and accuracy of the student learning situation feedback. The correlation information collection on the online education platform can realize real-time data information collection and acquisition, avoid data delay, ensure the timeliness and accuracy of the acquired data information, and MongoDB as a cloud database provides high availability and scalability, can effectively store and manage a large amount of student data, and the JSON format storage is convenient for data reading and processing, provides convenience for the storage of the second acquired data information and the third acquired data information, and also facilitates the student multi-source learning data set determination of the fusion processing module. Through the integrity analysis and processing of the original data set, the integrity of the student multi-source learning data set is ensured, the analysis deviation caused by data loss is avoided, and the reliability of the parent terminal applet system is enhanced.

[0077] In one embodiment of the application, the fusion processing module uses a data fusion method to process the student multi-source learning data set, including:

[0078] The student multi-source learning data set is subjected to data cleaning to obtain a first processed data set.

[0079] The data cleaning is to determine the abnormal value or inconsistent item data in the student multi-source learning data set, and to process the abnormal value or inconsistent item data in the student multi-source learning data set to determine the cleaned student multi-source learning data set and obtain the first processed data set.

[0080] The data format of the first processed data set is analyzed, and standardized processing is performed on the data format to obtain a second processed data set.

[0081] When the data format is standardized, the different format data in the first processed data set is converted into a unified standard format through normalization processing. For example, the first acquired data information (video viewing education situation information) in JSON format is parsed using a Python script, key fields are extracted, and the key fields are converted into a unified data model. The second acquired data information (exercise completion situation information) in CSV format is read using a Pandas library, the correct rate is calculated and a vector is generated, and then the timestamp alignment processing is performed. The third collected data information (test situation information) in XML format is parsed by an XSLT script, the score situation is extracted, and is normalized to the 0-1 interval to form a record form.

[0082] Fusion processing is performed in the second processed data set to determine the integrated data information.

[0083] In the second processing data set, the data information in the second processing data set is integrated to generate a unified data structure to obtain integrated data information. For example, the Apache Kafka stream processing pipeline is used to associate the standardized first acquisition data information, the standardized second acquisition data information, and the standardized third acquisition data information based on the user ID and the timestamp to generate a unified time series data set to obtain the integrated data information. Here, an example of the format of the data set of the integrated data information is as follows: {“user_id”: 789, “timestamp”: “2025-09-11T12:00:00Z”, “video_duration”: 38.5, “exercise_accuracy”: 0.89, “test_score”: 0.9}.

[0084] “video_duration”:38.5,“exercise_accuracy”:0.89,“test_score”:0.9}。

[0085] The standardized processing is performed on the integrated data information to obtain student multi-source learning processing data.

[0086] In the standardized processing of the integrated data information, principal component analysis is used to reduce the dimensionality of the integrated data information, calculate the index contribution rate, extract the main feature vector, and eliminate the dimensional difference through Z-score standardization to obtain the student multi-source learning processing data. The student multi-source learning processing data is stored in the Elasticsearch index to support subsequent query and analysis.

[0087] The above-mentioned data cleaning of the student multi-source learning data set discovers abnormal values or inconsistent item data in the student multi-source learning data set, guarantees the quality of the student multi-source learning data set, avoids misleading the behavior analysis module, and affects the accuracy of the student learning analysis data. The standardized processing of the data format makes the student multi-source learning data more easily processed and analyzed, reduces the complexity and conversion cost in the data processing process, provides protection for the subsequent analysis and processing of the student multi-source learning processing data, enables efficient data integration and analysis in the second processing data set, improves the efficiency of data processing, and more quickly generates a unified time series data set. In addition, the standardized processing of the integrated data information removes redundant information and extracts the most valuable features to improve the efficiency and accuracy of data analysis. Moreover, the Z-score standardization eliminates the dimensional difference to compare and analyze the data of different indicators under the same dimension, avoids analysis deviation caused by different dimensions, and improves the comparability and accuracy of the data.

[0088] As Figure 2As shown, the present application provides an embodiment, the behavior analysis module according to the student multi-source learning processing data for learning behavior analysis includes: behavior analysis unit and analysis diagnosis unit;

[0089] As shown, the behavior analysis unit according to the student multi-source learning processing data for learning behavior analysis includes the following steps: Figure 3

[0090] A1, based on the student multi-source learning processing data analysis and calculation of the student's knowledge point mastery rate, get the first learning analysis data.

[0091] Among them, when analyzing and calculating the student's knowledge point mastery rate, the target data information analysis is carried out from the student multi-source learning processing data, the course video completion degree of the student watching the video education is determined, and the submission frequency and test score of the exercise are obtained. The weighted average method is used to calculate the knowledge point mastery rate according to the course video completion degree, and the submission frequency and test score of the exercise are obtained, and the first learning analysis data is obtained. Here, the weights of course video completion degree, exercise submission frequency and test score are determined according to the course difficulty, for example, the weight of course video completion degree is 0.3, the weight of exercise submission frequency is 0.5, and the weight of test score is 0.2.

[0092] A2, analyze the student multi-source learning processing data, determine the student's learning time and homework submission frequency, and perform learning attitude clustering analysis according to the learning time and homework submission frequency, determine the student's learning attitude, and obtain the second analysis data.

[0093] Among them, when the learning attitude clustering analysis is performed according to the learning time and homework submission frequency, the data points of the students are determined according to the learning time and homework submission frequency of the students, and the distances between the data points and the clustering centers are calculated respectively, then the attribution of the data points is determined according to the size value of the distance, so as to determine the attribution category of the learning attitude of the students, and the learning attitude of the students is clear, and the second analysis data is obtained. Here, the student learning attitude is divided into three categories, respectively, positive learning attitude, general learning attitude and negative learning attitude.

[0094] A3, the first learning analysis data is combined with the learning time to predict the student learning behavior, and the third analysis data is obtained.

[0095] Among them, when the first learning analysis data is combined with the learning time to predict the student learning behavior, the learning behavior monitoring probability prediction analysis is performed based on the knowledge point mastery rate and the learning time through the logistic regression model, the possibility of the student's persistent learning is determined, and the third analysis data is obtained.

[0096] ​The technical solution realizes more in-depth learning analysis of students through the behavior analysis module, avoids the deviation of the knowledge point mastery rate of students caused by single data results, and the distortion of the learning behavior analysis of students, and the weight is determined according to the difficulty of the course, so that the evaluation result is more targeted and adaptive, and the learning situation of students is more accurately reflected. The learning attitude clustering analysis according to the learning duration and the homework submission frequency can quickly identify the learning behavior mode of students, avoid the deviation of subjective judgment, and objectively reflect the learning attitude category of students. The first learning analysis data is combined with the learning duration to predict the learning behavior of students, so that potential learning problems can be found in advance, and the basis for timely intervention is provided, so that students with low learning persistence probability can be provided with more learning support and incentive measures to help students improve the learning persistence probability and ensure the learning of students.

[0097] The learning diagnosis of the student by the analysis diagnosis unit according to the first learning analysis data includes the following steps:

[0098] The first learning analysis data is judged in combination with the first threshold value, and a first judgment result is obtained.

[0099] The first threshold value is a preset analysis threshold value for the knowledge point mastery rate. When the first learning analysis data is judged in combination with the first threshold value, the numerical size relationship between the first threshold value and the knowledge point mastery rate is analyzed, and the first judgment result is determined.

[0100] Error answering data is obtained according to the first judgment result, and the scale of the error answering data is judged in combination with a second threshold value, and a second judgment result is obtained.

[0101] When the first threshold value is greater than the knowledge point mastery rate, the error answering data is obtained from the exercise and test questions in the student multi-source learning processing data according to the first judgment result. When the first threshold value is not greater than the knowledge point mastery rate, the error answering data does not need to be obtained and analyzed. When the scale of the error answering data is judged in combination with the second threshold value, the number of error answering data is counted, the scale of the error answering data is determined, and then the size relationship between the scale of the error answering data and the second threshold value is compared to obtain the second judgment result. Here, the second threshold value is a threshold value set for comparing the scale of the error answering data.

[0102] The error answering data is expanded according to the second judgment result, and perfect data information is obtained.

[0103] When the second judgment result is that the scale of the error answering data is less than the second threshold value, the error answering data is clustered and analyzed, similar error records are merged, the scale of the error answering data is expanded, the expanded error answering data is determined, and the perfect data information is obtained.

[0104] The decision tree algorithm model is used to analyze the error type based on the improved data information, and the error type distribution information is determined.

[0105] The decision tree algorithm model is pre-constructed for error type identification analysis, and then obtained through optimization training. When the decision tree algorithm model is used to analyze the error type based on the improved data information, the error answering data is divided according to the root node of the decision tree, the application deficiency of the knowledge points is analyzed, the error type of the error answering data is determined, and thus the error type distribution information is obtained.

[0106] Based on the error type distribution information, the association between the error type and the knowledge point is analyzed, and the associated knowledge points are determined, and thus the weak knowledge point set is obtained.

[0107] When the association between the error type and the knowledge point is analyzed based on the error type distribution information, the high-frequency error type is extracted from the error type distribution, and the association between the error type and the knowledge point is analyzed and calculated based on the high-frequency error type. Then, based on the association, the error associated knowledge point is determined, the knowledge point error association set is obtained, and then the association strength analysis data is obtained by combining the specific knowledge point to analyze the knowledge point error association set. The association strength analysis data is analyzed and judged based on the third threshold value. When the association strength analysis data is higher than the third threshold value, the related knowledge points are extracted based on the specific knowledge point in the knowledge point error association set, and the weak knowledge point set is generated.

[0108] The family education guidance suggestion is generated according to the weak knowledge point set.

[0109] In the generation of family education guidance suggestions, the weak knowledge points in the weak knowledge point set are prioritized, the importance of the weak knowledge points and the error frequency of the students in the practice questions and test questions are determined, the weak knowledge points are arranged according to the priority determination rule, the priority order of the weak knowledge points is obtained, and then the learning plan of the weak knowledge points is made combined with the learning attitude of the students to determine the learning suggestions. For example, for students with a learning attitude of "positive", the first preset number of weak knowledge points are obtained according to the priority order of the weak knowledge points according to the priority level, and reflection and summary are made based on the obtained weak knowledge points and a large number of similar question training is made on the same weak knowledge point. Here, the first preset number is usually a ratio data, such as 60%. For students with a learning attitude of "general", the second preset number of weak knowledge points are obtained according to the priority order of the weak knowledge points according to the priority level, and reflection and summary are made based on the obtained weak knowledge points and similar question application exercises are made based on the same weak knowledge point. Here, the second preset number is less than the first preset number, such as 40%. For students with a learning attitude of "negative", the third preset number of weak knowledge points are obtained according to the priority order of the weak knowledge points according to the priority level, and application is found in life scenes for the obtained weak knowledge points, so as to guide learning in life scenes and improve the students' sense of achievement and mobilize the students' enthusiasm. Here, the third preset number is less than the second preset number, such as 10%. At the same time, the personality habits of the parents are obtained combined with the parent portrait information, and the family education guidance suggestions are made for the parents combined with the learning attitude of the students according to the personality habits of the parents, for example, for emotional parents, it is suggested that the parents give the students independent family learning space, and for stable parents, it is suggested that the parents accompany the students for family learning when the learning attitude of the students is negative or general.

[0110] The learning deficiencies of the student in the learning process are determined through the learning diagnosis of the student according to the first learning analysis data, so that the parents can quickly understand the learning state of the student, and the decision tree algorithm model can be ensured to have sufficient data information for analysis according to the second judgment result, while the attribute characteristics of the error answering data are not lost, the accuracy of the error type identification is improved, and the decision tree algorithm model can efficiently and accurately determine the error type for a large amount of and complex error answering data, thereby effectively improving the efficiency of the diagnosis and analysis. The correlation between the error type and the knowledge point is analyzed based on the error type distribution information, so that the root of the error answering is determined, the knowledge point that the student does not master in the learning process is determined, the weak knowledge point of the student is accurately positioned, the weak knowledge point set of the student is determined, and then the family education guidance suggestion is generated according to the weak knowledge point set, which not only provides a clear target for personalized learning intervention, but also provides a reference for the family education of the parents, thereby realizing good connection and cooperation between the family education and the school education, and providing stronger support and guarantee for the overall development of the student.

[0111] When the family education guidance suggestion is generated according to the weak knowledge point set, the family learning suggestion is not only provided for the student, but also the family education guidance suggestion is provided for the parents, the effect of the family education is improved, the student can improve the learning deficiencies in the family learning, and the relationship between the parents and the student is also guaranteed, so that the learning mood of the student is not affected by the parents. Moreover, when the family education guidance suggestion is provided for multiple students, the priority of the weak knowledge point in the weak knowledge point set is analyzed, the priority order of the weak knowledge point is determined, the family learning is appropriately combined with the learning attitude of the student, the learning pressure of the student is reduced, the student is prevented from having psychological problems due to the blind pursuit of comprehensive mastery, the family learning is appropriately combined with the learning state of the student, the effectiveness of the family learning is guaranteed, the growth space of the student is allowed to exist, the physical and mental health of the student is guaranteed, and when the family education guidance suggestion is provided for the parents, the appropriate learning space of the student is given, which helps the student to form good learning habits, avoids the influence of the learning mood of the student by the parents, and guarantees the family atmosphere.

[0112] In one embodiment of the present application, as shown in Figure 4 The report generation module generates a learning situation report text based on the student learning analysis data and the parent portrait information through a natural language processing model, including:

[0113] B1, the student learning analysis data is matched with a structured template to determine a structured target template.

[0114] Among them, the students correspond to different structured templates of different learning categories. Here, the learning categories include: academic learning, extracurricular development learning, etc.

[0115] B2, based on the structured target template, the learning situation initial report text information is generated according to the student learning analysis data through the natural language processing model.

[0116] Among them, the learning situation initial report text information is a concise version of the report text. When generating the learning situation initial report text information based on the structured target template through the natural language processing model according to the student learning analysis data, the key information of the student learning analysis data is parsed to determine the key information of the student learning analysis data, and the key information of the student learning analysis data is standardized according to the structured target template to obtain the key standardized information of the student learning analysis data; The natural language processing model is used to generate a statement based on the key standardized information of the student learning analysis data; The expression sentences are combined to obtain the learning situation initial report text information.

[0117] B3, adjust the style of the learning situation initial report text information according to the learning situation initial report text information, and obtain the learning situation adjusted report text information.

[0118] Among them, when adjusting the style of the learning situation initial report text information according to the learning situation initial report text information, the parent characteristics are determined according to the parent portrait information, and the expression adjustment and semantic enrichment are performed on the learning situation initial report text information according to the parent characteristics, and then the learning situation adjusted report text information is obtained. Here, the parent characteristics refer to the understanding ability of the parents to the report, the personality habits of the parents, etc., for example: for the emotional parents, the report is made by using the relatively implicit expression that the parents can understand.

[0119] B4, optimize the learning situation adjusted report text information, and obtain the final learning situation report text information.

[0120] Among them, when optimizing the learning situation adjusted report text information, the sentence coherence optimization and the ambiguity description optimization are performed on the learning situation adjusted report text information.

[0121] The learning situation report text can be efficiently and quickly obtained through the report generation module, and feedback can be timely given to the learning situation of the student, so that the parents can timely carry out family education according to the learning situation of the student, and thus good connection and cooperation of family education and school education are realized, and more powerful support and guarantee are provided for the all-round development of the student. The learning object of the student can be adapted by performing the structured template matching on the student learning analysis data, so that when the learning situation initial report text information is generated based on the structured target template by the natural language processing model according to the student learning analysis data, the learning situation initial report text information can more accurately reflect the learning situation of the student. Moreover, the style of the learning situation initial report text information is adjusted in combination with the parent portrait information, so that the learning situation adjustment report text information is easy for the report object to understand, and the problem that the parents cannot understand the report information and thus cannot master the learning situation of the student is effectively solved, the flexibility of the parent end small program system is improved, and the parents can effectively obtain the learning situation of the student from the parent end small program. In addition, the rationality and accuracy of the learning situation report text information are further ensured by optimizing the learning situation adjustment report text information.

[0122] In one embodiment of the present application, the style of the learning situation initial report text information is adjusted in combination with the parent portrait information, including:

[0123] The target parent portrait information is determined through the parent portrait recommendation.

[0124] In the parent end small program system, the number of parent portrait information can be one, two or more than two. When the style of the learning situation initial report text information is adjusted in combination with the parent portrait information, if the number of parent portrait information is one, the parent portrait information is the target parent portrait information, if the number of parent portrait information is two or more than two, the target parent portrait information is determined according to the priority of the parent portrait information, or the report object setting information is obtained, and then the report object setting information is screened in the parent portrait information according to the setting, so as to determine the target parent portrait information.

[0125] The cognitive level of the parent is analyzed according to the target parent portrait information, and the first feature of the parent is obtained.

[0126] When the cognitive level of the parent is analyzed according to the target parent portrait information, the understanding ability of the target parent to information and the ability level such as the focus on the student are analyzed based on the parent portrait information, so as to obtain the first feature of the parent.

[0127] The personality and habit of the parent are analyzed according to the target parent portrait information, and the second feature of the parent is obtained.

[0128] In the process of analyzing the character and habits of the parents according to the target parent portrait information, the character and habit information of the parents is extracted from the target parent portrait information to determine the character features of the parents, and a second feature of the parents is obtained. Here, the character and habit information of the parents is obtained by acquiring the reaction information of the parents through some scene test questions when the parent portrait information of the parents is acquired, and the character and habit information is analyzed according to the reaction information. The character and habit information includes emotional type, stable type, and indifferent type. For example, the emotional type is prone to impulsive after seeing sensitive information, and no matter what information the sensitive information represents in the context, the student is immediately criticized and educated. The stable type can communicate with the student in a peaceful manner on any matter. The indifferent type does not care much about the student, and the reaction to any situation is relatively slow.

[0129] According to the second feature of the parents, the initial report text information of the learning situation is adjusted to obtain first adjusted report text information of the learning situation.

[0130] In the process of adjusting the expression according to the second feature of the parents for the initial report text information of the learning situation, the reaction of the parents is predicted according to the initial report text information of the learning situation combined with the second feature of the parents to obtain the reaction prediction information of the parents, and it is determined whether the student will be misunderstood according to the reaction prediction information of the parents. When the student will be misunderstood, the sentence that intensifies the relationship between the parents and the student is determined in the initial report text information of the learning situation, the target sentence is locked, and the expression is adjusted for the target sentence. For example, the target sentence is "XX student's performance in the learning stage has dropped seriously", for the emotional type parents, the target sentence is adjusted to "XX student's performance in the learning stage has dropped seriously" to "XX student's performance in the learning stage has dropped seriously". In the learning process of the student in this stage, some challenges are encountered, and the performance fluctuates. For the indifferent type parents, the target sentence is adjusted to "XX student's performance in the learning stage has dropped seriously" to "XX student's performance in the learning stage has dropped seriously. In the learning process of the student in this stage, the performance has dropped from 90 points to 60 points, which has reached a very dangerous stage. If this continues, he / she will miss important learning opportunities and affect future development." The target sentence is adjusted to a more indirect or implicit expression to obtain the first adjusted report text information of the learning situation.

[0131] According to the first feature of the parents, the semantic richness of the first adjusted report text information of the learning situation is performed to obtain the second adjusted report text information of the learning situation.

[0132] Wherein, when the first adjustment report text information for the learning situation is semantically enriched according to the first characteristic of the parents, the first adjustment report text information for the learning situation is expanded in sentences in combination with the first characteristic of the parents, examples adapted to the first characteristic of the parents are used to assist the expression, the semantic expression of the sentences is enriched, so that the target parents can understand the meaning expressed through the second adjustment report text information for the learning situation, and clearly see the situation of the students in the information they focus on.

[0133] The above target parent portrait information determined through the parent portrait recommendation makes the report targeted, avoids the influence of report information not adapted to the parents on the report effect, and ensures that the parents can understand the learning situation of the students. The cognitive level analysis and personality habit analysis of the parents according to the target parent portrait information can understand the characteristics of the parents, improve the applicability of the learning situation adjustment report text information, ensure that the parents can understand the feedback information of the report information, and can also meet the preferences of the parents, can directly and efficiently and clearly see the situation of the students in the information they focus on, improve the satisfaction of the parents to the parent terminal program system, and can also reduce the possibility of misunderstanding of the report information by the parents, avoid the intensification of the relationship between the parents and the students, reduce the probability of emotionalization of the parents, and protect the physical and mental health of the students.

[0134] In one embodiment of the present application, when optimizing the learning situation adjustment report text information, the learning situation adjustment report text information is analyzed and optimized for sentence coherence, to obtain learning situation adjustment report text first optimization information; the completeness of the learning situation adjustment report text first optimization information is analyzed and judged, and when the learning situation adjustment report text first optimization information is complete, the learning situation adjustment report text first optimization information is sequentially analyzed for ambiguity description, the content with ambiguity description is locked, the content with ambiguity description is corrected through semantic analysis, the learning situation adjustment report text second optimization information is determined, and the final learning situation report text information is obtained.

[0135] Wherein, when the sentence coherence analysis and optimization of the learning situation adjustment report text information is performed, a sequence generation model is used to optimize the coherence of the sentence, the learning situation adjustment report text information is disassembled to obtain a structured sequence of the learning situation adjustment report text information, a connection analysis is performed on the structured sequence to determine whether there is a connection expression between the sentences, if there is a connection expression, it is determined whether the current connection expression is appropriate, if appropriate, optimization is not required, if inappropriate, a connection expression update is performed on the current connection expression, a new connection expression is determined, and the current connection expression is replaced with the new connection expression. If there is no connection expression, a connection word is determined according to the logical relationship between the sentences, and then a connection expression is determined based on the connection word, so that the sentence connection is performed according to the connection expression, and then the first optimization information of the learning situation adjustment report text is obtained.

[0136] When the content with ambiguous description is corrected through semantic analysis, the correct meaning of the content with ambiguous description is determined, the correct expression keyword is determined based on the correct meaning, and then the expression auxiliary information is generated according to the correct expression keyword, the content with ambiguous description is supplemented by using the expression auxiliary information, so that the correction of the content with ambiguous description is realized, and it is ensured that the content with ambiguous description can correctly express the meaning intended to be expressed.

[0137] The above-mentioned optimization of the learning situation adjustment report text information not only ensures the coherence of the sentence expression, avoids the influence on the parents, reduces the expression ability of the parents, but also avoids the misunderstanding of the parents caused by the ambiguity of the learning situation adjustment report text information, ensures the accuracy of the report, and enables the parents to accurately understand the learning situation of the students.

[0138] In one embodiment of the present application, before the report generation module generates the learning situation report text based on the student learning analysis data combined with the parent portrait information, the student learning analysis data is optimized and set, including:

[0139] Identifying and analyzing the weaknesses of the student to determine the type of weaknesses of the student;

[0140] The type of weaknesses of the student includes: short board of subject knowledge, defect of skill application or lack of learning habit. When the weaknesses of the student are identified and analyzed, a convolutional neural network model is used to extract and classify features from the multi-source learning processing data of the student, to determine the classification result and the weakness significance degree index data, and to analyze the weakness features according to the classification result and the weakness significance degree index data, so as to determine the type of weaknesses of the student. Here, the weakness significance degree index data is calculated based on the error rate, time deviation and historical trend.

[0141] The psychological sensitivity assessment engine is started based on the student weak item type, the psychological sensitivity of the student to the student weak item type is obtained, and a psychological sensitivity analysis result is determined.

[0142] In the process of starting the psychological sensitivity assessment engine, a multi-modal emotion inference model is used to analyze the psychological sensitivity in combination with the behavior trajectory data of the student, the avoidance tendency degree of the student to the fact that the student weak item type is known by others is analyzed based on the psychological sensitivity, the psychological sensitivity performance of the student to the student weak item type is determined, the psychological sensitivity of the student to the student weak item type is obtained, the psychological sensitivity of the student to the student weak item type is normalized, the psychological sensitivity of the student is converted to the range of [0, 1], normalized data is obtained, and it is determined whether the student weak item type needs to be hidden according to the normalized data, and a psychological sensitivity analysis result is obtained. Here, the behavior trajectory data is determined according to the performance of the student in the learning process based on the student multi-source learning data set, including: learning task click delay, error repetition mode, learning session interruption frequency, etc. The closer the normalized data is to 1, the more the student wants the student weak item type to be hidden.

[0143] According to the psychological sensitivity analysis result, the student learning analysis data is optimized, a hidden space is created, and student learning analysis optimization data is determined;

[0144] In the process of optimizing the student learning analysis data according to the psychological sensitivity analysis result, the student weak item type that the student wants to hide is determined according to the psychological sensitivity analysis result, the student learning analysis data associated with the student weak item type that the student wants to hide is filtered and extracted to obtain target data information, the hidden space is obtained according to the target data information, and the remaining student learning analysis data is used as optimized student learning analysis data to determine the student learning analysis optimization data.

[0145] The self-improvement possibility prediction engine is used to predict the self-improvement possibility of the hidden space, and a self-improvement probability is determined.

[0146] The self-improvement possibility prediction engine is a long short-term memory network model. When predicting the self-improvement possibility of the hidden space, the historical progress data of the student and the learning behavior consistency index data of the student are determined according to the student multi-source learning processing data analysis, and the self-improvement possibility prediction engine is used to predict the self-improvement probability of the student in a preset short-term time window according to the historical progress data of the student and the learning behavior consistency index data of the student by using the long short-term memory network model. Here, self-improvement refers to autonomous learning of the student without external intervention. The self-improvement probability is the possibility of improving the weak item type of the student through autonomous learning without external intervention. The progress history data refers to parameter data that can reflect the recent changes in learning of the student, such as the recent score change slope, error correction rate, etc. The learning behavior consistency index data is calculated by the entropy value of the behavior sequence. The higher the self-improvement probability, the greater the self-improvement potential.

[0147] The hidden time of the hidden space is determined by the hidden time calculation algorithm according to the self-improvement probability and the psychological sensitivity to the weak item type of the student, and the hidden space is set according to the hidden time;

[0148] The hidden time calculation algorithm is a calculation method model based on a dynamic weighting function, which uses an adjustment coefficient to harmonize the base duration, wherein the adjustment coefficient is obtained according to the self-improvement probability and the psychological sensitivity to the weak item type of the student in combination with an adaptive weight coefficient, and the calculation formula is as follows:

[0149] t = 1 + m x A - n x B

[0150] In the above formula, t is the adjustment coefficient, m is the first adaptive weight coefficient, n is the second adaptive weight coefficient, A is the psychological sensitivity to the weak item type of the student, and B is the self-improvement probability. Here, the first adaptive weight coefficient and the second adaptive weight coefficient are dynamically adjusted according to the parent portrait information, so as to ensure that the hidden time is adapted to the educational style of the parents and avoid negative harm to the student due to misunderstanding of the parents in the process of family education.

[0151] Learning situation report text generation is performed based on the hidden time. When the hidden time is valid, learning situation report text is generated based on student learning analysis optimization data and parent portrait information by using a natural language processing model, and the learning situation report text information is annotated according to the hidden space. At the same time, the hidden space is supervised, and emergency reports are made according to the supervision results. When the hidden time is invalid, the hidden space is unhidden, and learning situation report text is generated based on the hidden space and parent portrait information by using a natural language processing model.

[0152] The hidden time is valid if the hidden statistical time of the hidden space is within the set hidden time, and the hidden time is invalid if the hidden statistical time of the hidden space exceeds the set hidden time. When the hidden space reports the text information of the learning situation for annotation, the annotation information is determined according to the student weak item type in the hidden space and the psychological sensitivity of the student to the student weak item type, so as to prompt the parents through the annotation and avoid hurting the psychology of the students in the process of family education. When the hidden space is supervised, the actual progress of the student is monitored, the actual progress data of the student is obtained, and whether the progress of the student is lower than the expectation is analyzed and judged according to the actual progress data of the student. If yes, the hidden space is terminated, and an emergency reporting mode is started based on the hidden space, and the learning situation reporting text is generated and reported based on the hidden space.

[0153] Further, after the hidden space is released or terminated, the psychological sensitivity evaluation engine and the self-improvement possibility prediction engine are updated, the psychological sensitivity evaluation engine and the self-improvement possibility prediction engine are optimized through gradient descent, the accuracy of the psychological sensitivity evaluation engine is improved, and the prediction accuracy of the self-improvement possibility prediction engine is improved, so that the psychological sensitivity evaluation engine and the self-improvement possibility prediction engine achieve the purpose of continuous self-adaptation.

[0154] The above-mentioned identification analysis of the weaknesses of the students makes it possible to determine the type of weaknesses of the students when enabling the hidden space, which helps the school and the parents to better understand the needs of the students. The psychological sensitivity of the students is determined by starting the psychological sensitivity evaluation engine based on the type of weaknesses of the students, especially the psychological attitude of the students to the type of weaknesses of the students, which avoids causing psychological damage to the students with psychological sensitivity in the education process. Moreover, the psychological sensitivity of the students to the type of weaknesses of the students is normalized for comparison and analysis, which facilitates the determination of the psychological sensitivity analysis result. The learning analysis data of the students is optimized according to the psychological sensitivity analysis result, which protects the privacy and mental health of the students in the hidden space, reduces the psychological pressure and negative emotions of the students caused by the disclosure of the type of weaknesses of the students, helps the students to learn in a safer environment, and also provides sufficient private space for the students to learn and explore independently, which is beneficial to cultivate the self-learning ability of the students and improve the learning level of the students. The self-improvement possibility prediction engine is used to predict the self-improvement possibility of the hidden space, determine the current ability level of the students, and provide data support for the determination of the hidden time, so as to ensure the rationality of the hidden time. The determination and setting of the hidden time make the hidden space have timeliness, avoid the omission of data information in the hidden space affecting the learning of the students, and enable the emergency reporting mode to give priority to reporting and feedback based on the hidden space when the hidden space is released or terminated, so as to timely educate and guide the students according to the hidden space, protect the learning content of the students, and the hidden time is not fixed, but is associated with the self-improvement possibility prediction engine and the self-improvement possibility prediction engine, which can be improved with time and the changes of the students, and better serve the optimization of the learning analysis data of the students, and provide protection for the education and mental health of the students.

[0155] In one embodiment of the present application, when the situation feedback module feeds back the learning situation of the students according to the learning situation report text information, the historical feedback information is updated and fed back, including:

[0156] The learning situation report text information is segmented and disassembled to obtain a plurality of information blocks.

[0157] In the segmentation and disassembly, the segmentation and disassembly are usually performed according to the reporting framework at the time of feedback. The reporting framework is obtained by individualization, and the learning situation report text information is block processed according to the information type in the reporting framework, and each block can be set according to the preference.

[0158] According to the information block, the integrity analysis is combined with the historical feedback information, the corresponding relationship is determined, the missing information is determined, and the integrity analysis result is obtained.

[0159] According to the information block, the integrity analysis is combined with the historical feedback information, the corresponding relationship is determined, the missing information is determined, and the integrity analysis result is obtained.

[0160] According to the integrity evaluation result, the corresponding information block is marked according to the corresponding relationship and the historical feedback information, the corresponding historical feedback information is obtained for the missing information, and the missing information is supplemented according to the corresponding historical feedback information, and the missing supplement information block is obtained.

[0161] According to the information block, the integrity analysis is combined with the historical feedback information, the corresponding relationship is determined, the missing information is determined, and the integrity analysis result is obtained.

[0162] Integrate the marked information block and the missing supplement information block, and perform personalized reporting based on the integration result.

[0163] According to the information block, the integrity analysis is combined with the historical feedback information, the corresponding relationship is determined, the missing information is determined, and the integrity analysis result is obtained.

[0164] The learning situation feedback module can feed back the learning situation report text information to the parents more clearly, so that the parents can understand the learning situation of the students, and thus can better carry out family education for the students and guarantee the learning effect of the students. The learning situation report text information is segmented and disassembled, so that the feedback is not limited to a specific style during personalized reporting, and the parents can adjust the content and order of each block according to their own needs to realize personalized feedback. The integrity of the learning situation report text information is ensured by analyzing the completeness according to the information block and historical feedback information, so as to avoid missing information affecting the unity of personalized feedback, and also to guarantee the comprehensiveness of the student learning situation feedback. The changes of the current feedback information from the historical feedback information are clearly reflected during personalized reporting by marking, so as to facilitate the reporting object to understand the learning situation of the students based on the feedback information.

[0165] In one embodiment of the present application, the situation feedback module further adjusts the learning situation report text information in real time, comprising:

[0166] The reporting object is identified and analyzed to determine whether the reporting object is consistent with the current parent portrait, and an identification analysis result is obtained.

[0167] The identification scanning unit is arranged in the situation feedback module, and when the situation feedback module feeds back the learning situation of the student according to the learning situation report text information, information of the reporting object is acquired, and the reporting object information is analyzed in combination with the parent portrait corresponding to the learning situation report text information to determine whether the reporting object is consistent with the current parent portrait, and an identification analysis result is obtained.

[0168] When the identification analysis result is that the reporting object is inconsistent with the current parent portrait, a feedback adjustment request instruction is generated.

[0169] When the identification analysis result is that the reporting object is consistent with the current parent portrait, the learning situation report text information does not need to be adjusted. The feedback adjustment request instruction contains the identification features of the reporting object, so that the parent portrait of the reporting object can be directly acquired when the reporting generation module adjusts the report according to the learning situation report text information.

[0170] The reporting generation module adjusts the report according to the learning situation report text information based on the feedback adjustment request, and obtains adjusted learning situation report text information.

[0171] The reporting generation module adjusts the report according to the learning situation report text information based on the feedback adjustment request, and obtains adjusted learning situation report text information.

[0172] The adjusted learning report text information is fed back to the student in the situation feedback module.

[0173] The adjusted learning report text information is fed back to the student in the situation feedback module.

[0174] The above-mentioned identification analysis for the reporting object ensures that the reporting object can understand the learning situation reflected by the learning report text information, avoids the situation that the reporting object cannot read the learning report text information and cannot perform reasonable family education, enables the parents to better perform family education, guarantees the quality and effect of the family education, realizes good connection and cooperation between the family education and the school education, and provides stronger support and guarantee for the overall development of the students.

[0175] The first, second and third in the present application only refer to different application stages.

[0176] Other embodiments of the present disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the content disclosed herein. This application is intended to cover any variations, uses or adaptations of the present disclosure that follow, in general, the principles of the present disclosure and include other known ins and customary technical practices not specifically mentioned herein. The specification and examples are to be regarded as illustrative only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0177] It should be understood that the present disclosure is not limited to the precise structures described above and illustrated in the drawings and that various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is indicated by the appended claims.

Claims

1. A parent terminal applet system for recommending and reporting student learning based on a parent profile, characterized by, The method comprises the following steps: The data acquisition module is used to acquire multi-source learning data of students after the parent terminal app logs in, and obtain a student multi-source learning data set; meanwhile, a parent portrait recommendation is made for the parents of the students, and parent portrait information is determined; The fusion processing module is used to process the student multi-source learning data set by using a data fusion method, and obtain student multi-source learning processing data; The behavior analysis module is used to analyze the learning behavior of students based on the student multi-source learning processing data, and obtain student learning analysis data; The report generation module is used to generate a learning situation report text based on the student learning analysis data and the parent portrait information by using a natural language processing model, and obtain learning situation report text information; The situation feedback module is used to feed back the learning situation of students according to the learning situation report text information in the parent terminal app.

2. The parent applet system of claim 1, wherein, The data acquisition module acquires multi-source learning data of students, comprising: Determine the target student, and collect associated information based on the target student on the online education platform to obtain an original data set; The integrity of the original data set is analyzed and processed to obtain a student multi-source learning data set.

3. The parent applet system of claim 1, wherein, The fusion processing module processes the student multi-source learning data set by using a data fusion method, comprising: Data cleaning is performed on the student multi-source learning data set to obtain a first processing data set; The data format of the first processing data set is analyzed, and standardized processing is performed on the data format to obtain a second processing data set; Fusion processing is performed in the second processing data set to determine integrated data information; The integrated data information is standardized to obtain student multi-source learning processing data.

4. The parent applet system of claim 1, wherein, The behavior analysis module analyzes the learning behavior of students based on the student multi-source learning processing data, comprising: The behavior analysis unit: based on the student multi-source learning processing data, the knowledge point mastery rate of the student is analyzed and calculated to obtain first learning analysis data; the learning duration and homework submission frequency of the student are determined by analyzing the student multi-source learning processing data, and the learning attitude clustering analysis is performed according to the learning duration and homework submission frequency to determine the learning attitude of the student to obtain second analysis data; the first learning analysis data is combined with the learning duration to predict the learning behavior of the student to obtain third analysis data; The analysis and diagnosis unit is used to diagnose the learning of the student according to the first learning analysis data, comprising: combining the first threshold to judge the first learning analysis data to obtain a first judgment result; obtaining error answering data according to the first judgment result, and combining the second threshold to judge the scale of the error answering data to obtain a second judgment result; expanding the error answering data according to the second judgment result to obtain improved data information; using a decision tree algorithm model to analyze the error type of the improved data information to determine the error type distribution information; based on the error type distribution information, the association between the error type and the knowledge point is analyzed, and the associated knowledge points are determined to obtain a weak knowledge point set; generate a family education guidance suggestion according to the weak knowledge point set.

5. The parent applet system of claim 1, wherein, The report generation module generates a learning situation report text based on the student learning analysis data and the parent portrait information by using a natural language processing model, comprising: The structured template matching is performed on the student learning analysis data to determine a structured target template; Based on the structured target template, an initial report text information of the learning situation is generated from the student learning analysis data by a natural language processing model; The style of the initial report text information of the learning situation is adjusted based on the parent portrait information to obtain adjusted report text information of the learning situation; The adjusted report text information of the learning situation is optimized to obtain the final report text information of the learning situation.

6. The parent applet system of claim 5, wherein, The style adjustment of the initial report text information of the learning situation based on the parent portrait information includes: The target parent portrait information is determined based on the parent portrait recommendation; The first feature of the parent is obtained by analyzing the cognitive level of the parent based on the target parent portrait information; The second feature of the parent is obtained by analyzing the personality and habits of the parent based on the target parent portrait information; The first adjusted report text information of the learning situation is obtained by adjusting the expression of the initial report text information of the learning situation according to the second feature of the parent; The second adjusted report text information of the learning situation is obtained by enriching the semantics of the first adjusted report text information of the learning situation according to the first feature of the parent.

7. The parent applet system of claim 5, wherein, When optimizing the adjusted report text information of the learning situation, the sentence coherence of the adjusted report text information of the learning situation is analyzed and optimized to obtain the first optimized information of the adjusted report text information of the learning situation; the completeness of the first optimized information of the adjusted report text information of the learning situation is analyzed and judged, and when the first optimized information of the adjusted report text information of the learning situation is complete, the first optimized information of the adjusted report text information of the learning situation is sequentially analyzed for ambiguity description, the content with ambiguity description is locked, the content with ambiguity description is corrected through semantic analysis, the second optimized information of the adjusted report text information of the learning situation is determined, and the final report text information of the learning situation is obtained.

8. The parent applet system of claim 5, wherein, Before the report generation module generates the report text of the learning situation based on the student learning analysis data and the parent portrait information, the student learning analysis data is optimized and set, including: The student's weak points are identified and analyzed to determine the type of the student's weak points; Based on the type of the student's weak points, a psychological sensitivity evaluation engine is started to obtain the psychological sensitivity of the student to the type of the student's weak points, and a psychological sensitivity analysis result is determined; The student learning analysis data is optimized according to the psychological sensitivity analysis result, a hidden space is created, and student learning analysis optimization data is determined; The self-improvement probability of the hidden space is predicted by a self-improvement probability prediction engine to determine the self-improvement probability; The hidden time of the hidden space is determined based on the self-improvement probability and the psychological sensitivity to the type of the student's weak points by a hidden time calculation algorithm, and the hidden space is set according to the hidden time; The learning situation report text is generated based on the hidden time. When the hidden time is valid, the learning situation report text is generated based on the student learning analysis optimization data and the parent portrait information by the natural language processing model, and the learning situation report text information is annotated according to the hidden space. At the same time, the hidden space is supervised, and emergency reports are made according to the supervision results. When the hidden time is invalid, the hidden space is unhidden, and the learning situation report text is generated based on the hidden space and the parent portrait information by the natural language processing model.

9. The parent applet system of claim 1, wherein, When the situation feedback module feeds back the student learning situation according to the learning situation report text information, the historical feedback information is updated and fed back, including: The learning situation report text information is segmented and disassembled to obtain a plurality of information blocks; According to the information block, the historical feedback information is analyzed for completeness, the corresponding relationship is determined, the missing information is determined, and the completeness analysis result is obtained; According to the completeness evaluation result, the corresponding information block is marked according to the corresponding relationship and the historical feedback information, the corresponding historical feedback information is obtained for the missing information, and the missing information is supplemented according to the corresponding historical feedback information to obtain the missing supplement information block; Integrate the marked information block and the missing supplement information block, and make individualized report based on the integration result.

10. The parent applet system of claim 9, wherein, The situation feedback module also adjusts the learning situation report text information in real time, including: Identify and analyze the report object to determine whether the report object is consistent with the current parent portrait, and obtain an identification analysis result; When the identification analysis result is that the report object is inconsistent with the current parent portrait, a feedback adjustment request instruction is generated; Based on the feedback adjustment request, the learning situation report text information is adjusted by the report generation module to obtain adjusted learning situation report text information; The adjusted learning situation report text information is fed back to the student learning situation in the situation feedback module.

Images