A learning situation analysis method, system, terminal and medium based on a large language model

Abstract

The application relates to a learning situation analysis method and system based on a large language model, a terminal and a medium, and relates to the field of education system technology. The method comprises the following steps: collecting learning data of students, the learning data comprising structured learning data and unstructured learning data, the structured learning data comprising test scores and homework completion conditions, and the unstructured learning data comprising classroom behavior data, learning process data, student self-evaluation and mutual evaluation data and teacher teaching records; pre-processing the learning data to generate standardized data that can be used for analysis; receiving input interaction information, generating analysis requirement parameters, and the interaction information comprising analysis requirements of teachers or feedback information of students; based on a large language model, performing semantic understanding and deep analysis on the standardized data and the analysis requirement parameters to obtain learning situation analysis results; and displaying the learning situation analysis results in a visual manner. The application has the effect of improving teaching quality and meeting actual teaching requirements.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of educational system technology, and particularly to a learning situation analysis method, system, terminal and medium based on a large language model. BACKGROUND

[0002] In the modern education system, a learning situation analysis system is a key technical support for improving teaching quality and meeting actual teaching needs.

[0003] In related technologies, a learning situation analysis system is based on a large language model technology, processes student learning data, matches a teaching standard database, generates a learning situation diagnosis report and teaching improvement suggestions, and finally forms a visual analysis chart.

[0004] For the above related technologies, the single analysis method and fixed index system have the problems of insufficient learning situation analysis accuracy and pertinence, which leads to the disconnection between the analysis results and actual teaching needs, and cannot provide effective support for teaching decisions. SUMMARY

[0005] In order to improve teaching quality and meet actual teaching needs, the present application provides a learning situation analysis method, system, terminal and medium based on a large language model.

[0006] In a first aspect, the present application provides a learning situation analysis method based on a large language model, which adopts the following technical solution:

[0007] A learning situation analysis method based on a large language model, comprising:

[0008] Collecting learning data of students, the learning data including structured learning data and unstructured learning data, the structured learning data including examination scores and homework completion, and the unstructured learning data including classroom behavior data, learning process data, student self-evaluation and mutual evaluation data, and teacher teaching records;

[0009] Preprocessing the learning data to generate standardized data that can be used for analysis;

[0010] Receiving input interaction information, generating analysis requirement parameters, the interaction information including teacher analysis requirements or student feedback information;

[0011] Based on a large language model, performing semantic understanding and deep analysis on the standardized data and the analysis requirement parameters to obtain learning situation analysis results;

[0012] Displaying the learning situation analysis results in a visual manner.

[0013] By adopting the technical scheme, the structured and unstructured learning data of students are fused, the deep understanding and personalized analysis of learning characteristics are realized by using a large language model, the analysis results are dynamically generated in combination with the teaching needs of teachers and the feedback of students, and the analysis results are presented in a multi-dimensional visualization manner, so that the comprehensiveness, accuracy and interactivity of learning analysis are effectively improved, and the scientificity of teaching decision and the pertinence of learning guidance are significantly enhanced.

[0014] Optionally, the consistency of the classroom behavior data and the learning process data is detected;

[0015] If the consistency does not meet a preset standard, when the classroom behavior data is higher than a first data threshold and the learning process data is lower than a second data threshold, it is determined that there is a logical conflict;

[0016] In response to the logical conflict, effective participation features in the monitoring video data in a conflict time period are analyzed to obtain an effectiveness score, the effective participation features including an angle between a face orientation and a blackboard or a teacher, a relevance between a body movement and a learning task, and an effective speech in a group interaction;

[0017] If the effectiveness score exceeds an effectiveness threshold, a first weight coefficient of the classroom behavior data is increased, and a second weight coefficient of the learning process data is decreased;

[0018] If the effectiveness score does not exceed the effectiveness threshold, a third weight coefficient of the classroom behavior data is decreased, and a fourth weight coefficient of the learning process data is increased;

[0019] The classroom behavior data and the learning process data are weighted and fused according to the first weight coefficient, the second weight coefficient, the third weight coefficient and the fourth weight coefficient to generate the standardized data.

[0020] By adopting the technical scheme, the consistency detection mechanism of the classroom behavior data and the learning process data is introduced, and the dynamic weight adjustment is performed in combination with the effective participation features in the monitoring video, so that the real learning state of students can be more accurately reflected, the intelligentization and adaptive ability of the model are improved, the generation of the standardized data is more objective and targeted, and the accuracy and reliability of the subsequent analysis results are improved.

[0021] Optionally, state features of students are collected, the state features including a body movement amplitude feature, a face orientation angle change feature and a group interaction state feature;

[0022] When the state features exceed a safety threshold, it is determined that there is an abnormal teaching event;

[0023] if the abnormal teaching event exists, freezing the classroom behavior data, marking the classroom behavior data as invalid data segment and removing;

[0024] taking valid data before the generation time of the invalid data segment as regular data;

[0025] based on the regular data, generating and calculating missing data and confidence;

[0026] if the confidence exceeds a threshold, adding the missing data to the classroom behavior data;

[0027] if the confidence does not exceed the threshold, freezing the missing data.

[0028] By adopting the above technical solution, the student state characteristics are intelligently collected and analyzed, combined with the recognition of abnormal teaching events and the data validity judgment mechanism, the abnormal data can be eliminated and intelligently completed and confidence evaluated based on historical regular data, the accuracy of classroom behavior data and the robustness of analysis model are improved, thereby ensuring the reliability and practicality of the learning situation analysis result, and enhancing the adaptability and intelligent level of the system in complex teaching environment.

[0029] Optionally, a fairness index of the learning situation analysis result is detected, the fairness index including at least one of score distribution difference of different gender student groups, recommendation similarity of different academic foundation student groups, and attention deviation value of special needs student groups;

[0030] if any of the fairness indexes exceeds a deviation threshold, a feature importance distribution is generated according to the fairness index, and the learning situation analysis result is recorded as a learning situation analysis result with potential deviation, the feature importance distribution containing influence degree scores of input features of the large language model on the final analysis result;

[0031] according to the feature importance distribution, a fairness constraint condition is dynamically configured in the large language model, and a corrected learning situation analysis result is regenerated based on the fairness constraint condition;

[0032] the learning situation analysis result with potential deviation and the corrected learning situation analysis result are simultaneously displayed in a visual interface.

[0033] By adopting the above technical solution, the fairness index detection mechanism is introduced, combined with feature importance analysis and dynamic constraint configuration, the potential deviation in the learning situation analysis result can be identified and corrected, the transparency and explainability of the analysis process are improved, the adaptability and fairness of the system to different student groups are enhanced, and more fair and reliable learning situation support is provided for teaching decision-making.

[0034] Optionally, a model accuracy change amount when the fairness constraint is applied is measured;

[0035] When the model accuracy change amount decreases by more than an accuracy threshold, a fairness-accuracy evaluation process is started, and key performance data of the large language model is output, including original model accuracy, model accuracy under the fairness constraint, and prediction difference between sensitive attribute groups;

[0036] According to the key performance data, the relationship between the fairness improvement degree and the model accuracy change amount is calculated, and multiple optional mediation schemes are generated, which contain different combinations of fairness and accuracy;

[0037] According to the mediation scheme, a bias parameter is selected, the updated constraint strength parameter is dynamically calculated, and the constraint strength parameter is fed back to the fairness constraint for adaptive adjustment. The bias parameter is 0, which represents accuracy priority, and the bias parameter is 1, which represents fairness priority.

[0038] By using the above technical solution, model performance detection and multi-objective optimization mechanism are introduced, which can dynamically adjust the fairness constraint strength while ensuring the overall accuracy of the model, realize the adaptive balance of fairness and accuracy, improve the flexibility and applicability of the system in different application scenarios, ensure that the learning situation analysis result takes into account fairness while maintaining high analysis reliability, and enhance the controllability and practical value of the model.

[0039] Optionally, at least one difference feature parameter between teachers and students is extracted, including knowledge point coverage difference, training intensity difference, or cognitive load difference;

[0040] Based on the difference feature parameter, a comprehensive conflict index is calculated;

[0041] If the comprehensive conflict index exceeds a comprehensive threshold, based on the learning situation analysis result, a first expected effect of a teacher suggestion and a second expected effect of a student feedback scheme are predicted respectively, and a coordination scheme is generated;

[0042] According to the coordination scheme, the display content and form of the coordination scheme are dynamically configured, and the learning situation analysis result is optimized.

[0043] By using the above technical solution, the knowledge point coverage, training intensity, and cognitive load difference between teachers and students are extracted, the coordination scheme is generated, and the display content and form are dynamically configured, which can effectively identify and alleviate the deviation in learning situation cognition between the two parties in teaching, improve the pertinence and adaptability of teaching intervention, and enhance the efficiency of teacher-student interaction and the intelligent level of teaching decision-making.

[0044] Optionally, when it is detected that the coordination scheme does not achieve the expected effect for N consecutive times, a traceability process is triggered;

[0045] Based on the traceability process, historical feature parameters in the coordination scheme are extracted, and a feature parameter of a dominant conflict is determined;

[0046] A conflict traceability report is generated according to the feature parameter, and the conflict traceability report includes a duration change curve of a key conflict parameter and a matching degree analysis of related teaching resources;

[0047] An alarm signal and the conflict traceability report are pushed to an administrator terminal, and an automatic coordination function on a student terminal and a teacher terminal is locked;

[0048] In response to receiving a manual confirmation instruction, the automatic coordination function on the student terminal and the teacher terminal is unlocked.

[0049] By adopting the above technical solution, the continuous monitoring and traceability mechanism of the coordination effect is introduced, which can automatically trigger traceability analysis when multiple coordinations are invalid, accurately identify the key factors of the dominant conflict, and generate a traceability report. Combined with manual intervention, the reliability and safety of system decision-making are improved, effectively preventing the repeated execution of invalid mediation, and enhancing the intelligent management level and abnormal processing capability of the teaching intervention system.

[0050] In a second aspect, the present application provides a learning situation analysis system based on a large language model, which adopts the following technical solution:

[0051] A learning situation analysis system based on a large language model, comprising:

[0052] An acquisition module for acquiring learning data and interaction information;

[0053] A memory for storing the program of the learning situation analysis method based on a large language model;

[0054] A processor, the program in the memory can be loaded and executed by the processor and implement the learning situation analysis method based on a large language model.

[0055] By adopting the above technical solution, the acquisition module collects student learning data in real time, the processor efficiently executes the learning situation analysis algorithm based on a large language model, and the memory continuously optimizes the analysis model and the data feature library, realizing intelligent processing of the whole process from data acquisition to learning situation diagnosis to teaching intervention. While improving the analysis accuracy and pertinence, the teaching decision-making efficiency is significantly enhanced, providing efficient and reliable technical support for precise teaching and individualized education in the smart education scenario.

[0056] In a third aspect, the present application provides an intelligent terminal, which adopts the following technical solution:

[0057] An intelligent terminal comprises a memory and a processor, and the memory stores a computer program capable of being loaded by the processor and executing any one of the above methods.

[0058] In a fourth aspect, the present application provides a computer storage medium capable of storing a corresponding program, having the characteristics of facilitating the improvement of teaching quality and meeting the actual teaching needs, and adopting the following technical solutions:

[0059] A computer-readable storage medium stores a computer program capable of being loaded by a processor and executing any one of the above learning situation analysis methods based on a large language model.

[0060] In summary, the present application includes at least one of the following beneficial technical effects:

[0061] Fusing students' structured and unstructured learning data, using a large language model to realize deep understanding and personalized analysis of learning situation characteristics, combining teachers' teaching needs and students' feedback to dynamically generate analysis results, and presenting through multi-dimensional visualization, effectively improving the comprehensiveness, accuracy and interactivity of learning situation analysis, significantly enhancing the scientificity of teaching decision-making and the pertinence of learning guidance;

[0062] Introducing a consistency detection mechanism for classroom behavior data and learning process data, and combining effective participation features in monitoring videos for dynamic weight adjustment, which can more accurately reflect students' real learning state, improve the intelligence and adaptability of the model, make the generation of standardized data more objective and targeted, and thus improve the accuracy and reliability of subsequent analysis results;

[0063] Introducing a fairness index detection mechanism, combining feature importance analysis and dynamic constraint configuration, which can identify and correct potential biases in learning situation analysis results, improve the transparency and explainability of the analysis process, enhance the adaptability and fairness of the system to different student groups while ensuring analysis accuracy, and provide more fair and reliable learning situation support for teaching decisions. BRIEF DESCRIPTION OF DRAWINGS

[0064] Figure 1 is a flowchart of a learning situation analysis method based on a large language model provided by an embodiment of the present application.

[0065] Figure 2 is a flowchart of a classroom learning state adaptive analysis method provided by an embodiment of the present application.

[0066] Figure 3 is a flowchart of a classroom behavior data intelligent repair method provided by an embodiment of the present application.

[0067] Figure 4is a flowchart of a fairness correction method for learning situation analysis provided by an embodiment of the present application.

[0068] Figure 5 is a flowchart of a fairness and accuracy adaptive balancing method provided by an embodiment of the present application.

[0069] Figure 6 is a flowchart of an intelligent mediation method for teacher-student teaching conflicts provided by an embodiment of the present application.

[0070] Figure 7 is a flowchart of a teaching mediation failure traceability intervention method provided by an embodiment of the present application.

[0071] Figure 8 is a structural diagram of a learning situation analysis system based on a large language model provided by an embodiment of the present application. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical scheme and advantages of the present application clearer, the following will further describe the present application in combination with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application. Figures 1 to 8

[0073] An embodiment of the present application discloses a learning situation analysis method based on a large language model. Referring to Figure 1 , the method comprises:

[0074] Step S101: Collecting learning data of students, the learning data including structured learning data and unstructured learning data, the structured learning data including examination results and homework completion, and the unstructured learning data including classroom behavior data, learning process data, student self-evaluation and mutual evaluation data, and teacher teaching records.

[0075] Learning data refers to a collection of data reflecting student learning behavior, knowledge mastery, and learning attitude, etc.

[0076] Structured learning data refers to data with fixed format and explicit fields.

[0077] Unstructured data refers to data with no fixed format and difficult to directly quantify and analyze.

[0078] For example, the historical examination results and homework submission records of students in the mathematics subject are obtained, and the classroom speaking frequency and posture change are collected by using a classroom camera.

[0079] Step S102: Preprocessing the learning data to generate standardized data that can be used for analysis.

[0080] ​Preprocessing refers to cleaning, converting and integrating the collected multi-source data.

[0081] Standardized data refers to the unified format data obtained by cleaning, normalizing and feature extraction on the collected original learning data.

[0082] For example, missing value filling and outlier removal are performed on structured data, and normalization method is used to standardize test scores and homework scores; unstructured data is segmented, stop word filtered, semantically annotated and sentiment analyzed, and key words and feature vectors are extracted to generate standardized data that can be used as model input.

[0083] Step S103: receiving input interaction information, generating analysis requirement parameters, the interaction information including the analysis requirements of teachers or the feedback information of students.

[0084] Analysis requirement parameters refer to structured control signals generated after parsing the original interaction information.

[0085] For example, the teacher inputs "analyze Zhang San's knowledge mastery in the part of physical mechanics" on the terminal interface, and through natural language processing technology, the analysis object identifier is identified as "Zhang San", the analysis dimension parameter is "the part of physical mechanics", and the corresponding analysis requirement parameter is generated; if the student feedbacks "I feel difficult to understand in circuit analysis", the system extracts the analysis object identifier as the current student, the analysis dimension parameter as "circuit analysis", and generates the corresponding analysis requirement parameter.

[0086] Step S104: based on the large language model, the standardized data and the analysis requirement parameters are semantically understood and deeply analyzed to obtain the learning situation analysis result.

[0087] Large language model refers to a language model with strong semantic understanding and reasoning ability, which can deeply analyze the semantic analysis of input standardized data and analysis requirement parameters.

[0088] Learning situation analysis result refers to the comprehensive evaluation conclusion output by the large language model after multi-dimensional analysis of standardized learning data and analysis requirement parameters.

[0089] For example, it is identified that a student performs well in algebra reasoning, but has frequent errors in geometric proof, so it is judged that there is a weak link in this knowledge point.

[0090] Step S105: visually display the learning situation analysis result.

[0091] Visual display refers to presenting the learning situation analysis result in the form of charts, graphs or figures.

[0092] Exemplarily, the knowledge graph is used to show the mastery correlation between different knowledge points of the student; the heat map is used to show the mastery degree distribution of the student on each knowledge point; and the line chart is used to show the learning trend change of the student.

[0093] By adopting the above technical solutions, the structured and unstructured learning data of the student are fused, the deep understanding and personalized analysis of the learning situation characteristics are realized by using the large language model, the analysis results are dynamically generated in combination with the teaching needs of the teacher and the feedback of the student, and the analysis results are presented in a multi-dimensional visualization manner, so that the comprehensiveness, accuracy and interactivity of the learning situation analysis are effectively improved, and the scientificity of the teaching decision and the pertinence of the learning guidance are significantly enhanced.

[0094] Embodiments of the present application disclose a classroom learning state adaptive analysis method. Referring to Figure 2 The method comprises the following steps:

[0095] Step S201: detecting the consistency of the classroom behavior data and the learning process data.

[0096] The classroom behavior data refers to the behavior characteristics of the student in the classroom collected by cameras, sensors and other devices, such as the attendance of the student and the classroom interaction performance.

[0097] The learning process data refers to the non-behavioral data generated by the student in the learning process, such as the learning duration and learning path on the online learning platform.

[0098] Exemplarily, the classroom behavior data of a student frequently speaking in a class and focusing his eyes on the blackboard is collected, and the learning process data of the student not browsing knowledge points for a long time on the learning platform is recorded, and the system determines whether the two are consistent.

[0099] Step S202: if the consistency does not meet the preset standard, when the classroom behavior data is higher than the first data threshold and the learning process data is lower than the second data threshold, it is determined that there is a logical conflict.

[0100] The first data threshold refers to the activity level for measuring the classroom behavior.

[0101] The second data threshold refers to the continuity or effectiveness for measuring the learning process.

[0102] The logical conflict refers to the significant difference between the learning state reflected by the classroom behavior data and the learning process data.

[0103] Exemplarily, the number of times of speaking in class and the time of focusing eyes on the blackboard of a student both exceed the first data threshold, but the number of times of clicking on knowledge points in the learning platform does not reach the second data threshold, so it is determined that there is a logical conflict between the classroom behavior data and the learning process data of the student.

[0104] Step S203: In response to the logical conflict, analyze the effective participation features in the monitoring video data in the conflict time period to obtain an effectiveness score, the effective participation features including an angle between a face orientation and a blackboard or a teacher, a relevance between a body action and a learning task, and an effective speech in a group interaction.

[0105] The effectiveness score refers to a real learning participation degree of a student in the conflict time period.

[0106] Through face recognition and posture detection based on the monitoring video data, a student position is located and a face orientation angle between the student and a blackboard or a teacher is tracked to determine a concentration situation, a relevance between a body action and a learning task is recognized, and a semantic feature of a speech content in a group discussion is analyzed by voice recognition to determine whether the speech content is around a learning theme, so as to comprehensively obtain an effective participation degree.

[0107] The effectiveness score is obtained by weighted calculation of the face orientation angle, the body action relevance, and the group speech effectiveness, wherein the face orientation weight is 0.4, the body action weight is 0.3, and the group speech weight is 0.3, each score is normalized, and a final effectiveness score is obtained by weighted summation to determine a real learning participation degree of a student.

[0108] For example, a video segment of a student in the conflict time period is called, it is analyzed that the face is always oriented to the teacher and the student speaks many times in the group discussion, and finally an effectiveness score of 75 is generated, indicating that the actual participation degree is high.

[0109] Step S204: If the effectiveness score exceeds an effectiveness threshold, a first weight coefficient of the classroom behavior data is increased, and a second weight coefficient of the learning process data is decreased.

[0110] When the effectiveness score exceeds the effectiveness threshold, it indicates that the actual learning participation degree of the student in the conflict time period is high, and the monitoring video data analysis result verifies the authenticity and concentration of the classroom behavior data of the student, so that increasing the weight of the classroom behavior data can more accurately reflect the learning state of the student, and relatively decreasing the weight of the learning process data can avoid misleading the analysis result due to a short-term low activity.

[0111] Step S205: If the effectiveness score does not exceed the effectiveness threshold, a third weight coefficient of the classroom behavior data is decreased, and a fourth weight coefficient of the learning process data is increased.

[0112] When the effectiveness score does not exceed the effectiveness threshold, it indicates that the actual learning participation of the student in the conflict time period is low, and analyzing the classroom behavior data may have the case of surface activity but insufficient substantive learning investment, therefore, the weight of the classroom behavior data is reduced to avoid misleading the analysis of the learning situation, and the weight of the learning process data is increased to more accurately reflect the real cognitive state and learning effect of the student, thereby ensuring the objectivity and reliability of the analysis result.

[0113] Step S206: The adjusted classroom behavior data and the learning process data are weighted and fused according to the first weight coefficient, the second weight coefficient, the third weight coefficient and the fourth weight coefficient to generate standardized data.

[0114] After the adjusted classroom behavior data and the learning process data are multiplied by the corresponding weight coefficients, the weighted sum is obtained, and then the normalized processing is performed to map it to a uniform numerical interval.

[0115] By adopting the above technical scheme, the consistency detection mechanism of the classroom behavior data and the learning process data is introduced, and the dynamic weight adjustment is combined with the effective participation features in the monitoring video, which can more accurately reflect the real learning state of the student, improve the intelligentization and adaptive ability of the model, make the generation of the standardized data more objective and targeted, and thus improve the accuracy and reliability of the subsequent analysis result.

[0116] The embodiment of the application discloses a kind of classroom behavior data intelligent repair method. Refer to Figure 3 The method comprises:

[0117] Step S301: collect the state features of students, and the state features include limb action amplitude features, face orientation angle change features and group interaction state features.

[0118] The state features refer to multi-dimensional features reflecting the behavior performance of students in the classroom.

[0119] Through the camera, the limb action of a student standing frequently and waving hands with large amplitude in a class is collected, combined with the angle change of its face orientation to the blackboard, and the behavior features of speaking in group discussion multiple times, the image recognition algorithm is used to analyze the limb action amplitude and frequency, the posture estimation model is used to calculate the angle change between the face orientation and the blackboard or the teacher, and the speech recognition and semantic analysis technology are combined to judge the speaking frequency and content relevance in group interaction.

[0120] Step S302: when the state features exceed the safety threshold, it is determined that there is an abnormal teaching event.

[0121] The abnormal teaching event refers to an abnormal behavior event occurring in the teaching process, such as emotional agitation or fighting of students.

[0122] By comparing the student state features in real time, when the amplitude of body movement exceeds 1.5 times of the set standard value, the face orientation angle is greater than 60 degrees and the duration is more than 30 seconds, the frequency of group interaction is more than 10 times per minute, or multiple people speak at the same time leading to voice overlap degree more than 40%, it is determined that the state feature is abnormal, and further identified as an abnormal teaching event that may exist interference with the teaching order.

[0123] For example, a student continuously and frequently taps the desk, stands and walks, and faces away from the teacher in a class. The amplitude of his body movement and the change of his face angle both exceed the set safety threshold. The system determines that there is an abnormal teaching event in this time period.

[0124] Step S303: If there is an abnormal teaching event, freeze the classroom behavior data, mark the classroom behavior data as invalid data segment and remove it.

[0125] Freezing classroom behavior data means suspending the use of classroom behavior data collected during the abnormal teaching event period.

[0126] Invalid data segment means marking the data in the abnormal teaching event period as abnormal data, which does not participate in subsequent analysis.

[0127] For example, the classroom speech frequency and eye trajectory of students collected during the abnormal teaching event are marked as invalid data segment and removed, so as to avoid interference with the subsequent analysis results.

[0128] Step S304: Take the valid data before the generation time of the invalid data segment as the regular data.

[0129] Regular data refers to the data set collected before the invalid data segment and conforming to the normal student state features.

[0130] For example, the classroom behavior data collected within 5 minutes before the invalid data segment is extracted as regular data, which is used for the generation and completion of missing data.

[0131] Step S305: Based on the regular data, generate and calculate the missing data and confidence.

[0132] Missing data refers to the classroom behavior data that is missing due to freezing of abnormal events.

[0133] Using regular data as historical input, time series prediction algorithm is used to model the trend of classroom behavior data, and then the classroom behavior data that should have appeared in the abnormal period is predicted according to the model output.

[0134] The confidence formula of the missing data is C=Cmax-k*t, wherein Cmax is a maximum confidence benchmark value, which is 0.95 by default, t is a data missing duration, and k is an attenuation coefficient, which is 0.1 / min by default.

[0135] For example, if an abnormal teaching event causes 2 minutes of missing behavior data, the confidence of the missing data is C=0.95-0.1*2=0.75.

[0136] Step S306: If the confidence exceeds the threshold value, the missing data is added to the classroom behavior data.

[0137] When the confidence of the missing data is higher than the threshold value, it indicates that the data has a high reliability based on historical regular data and can accurately reflect the learning state of the student in the abnormal time period. Therefore, the missing data is supplemented to the classroom behavior data, which can improve the data integrity and the reliability of the learning situation analysis result.

[0138] For example, the confidence of a certain segment of missing data is 0.75, which is higher than the threshold value 0.7, so it is included in the classroom behavior data set and used to generate the learning situation analysis result.

[0139] Step S307: If the confidence does not exceed the threshold value, the missing data is frozen.

[0140] When the confidence of the missing data is lower than the threshold value, the data is considered unreliable, and the missing data is frozen and does not participate in subsequent analysis.

[0141] For example, the confidence of a certain segment of missing data is only 0.6, which is lower than the threshold value 0.7, so the missing data is frozen and does not participate in the classroom behavior data analysis.

[0142] By using the above technical solution, the student state characteristics are intelligently collected and analyzed, and the identification of abnormal teaching events and the data validity judgment mechanism are combined. The abnormal data can be eliminated, and the historical regular data can be intelligently completed and confidence evaluated, which improves the accuracy of the classroom behavior data and the robustness of the analysis model, thereby ensuring the reliability and practicality of the learning situation analysis result, and enhancing the adaptability and intelligent level of the system in a complex teaching environment.

[0143] Embodiments of the present application disclose a learning situation analysis fairness correction method. Referring to Figure 4 The method comprises the following steps:

[0144] Step S401: Detect a fairness index of a learning situation analysis result, wherein the fairness index comprises at least one of a score distribution difference of different gender student groups, a recommendation similarity of different academic foundation student groups, and an attention deviation value of a special needs student group.

[0145] The fairness indicator is a quantitative indicator for measuring whether there is a systematic deviation in the learning profile analysis results among different student groups.

[0146] By statistically analyzing the data of different student groups in the learning profile analysis results, the score distribution difference of different gender student groups is calculated, the semantic similarity of learning suggestions obtained by students with different academic bases is compared, and the attention degree of special needs student groups in the analysis results is evaluated, thereby detecting whether there is a potential systematic deviation in the learning profile analysis results.

[0147] Step S402: If any fairness indicator exceeds the deviation threshold, the feature importance distribution is generated according to the fairness indicator, and the learning profile analysis result is recorded as a learning profile analysis result with potential deviation, and the feature importance distribution contains the influence degree score of the input features of the large language model on the final analysis result.

[0148] The feature importance distribution refers to the distribution formed by scoring the importance of each input feature affecting the learning profile analysis result, which is used to identify which features have a greater impact on the generation of deviation.

[0149] After detecting that the fairness indicator exceeds the deviation threshold, the influence degree of each input feature is analyzed based on the correlation between the input features and the output results of the large language model, and the contribution degree of different features to the analysis result is quantified, thereby generating the feature importance distribution for identifying the key features causing the deviation.

[0150] For example, if it is detected that the gender score difference exceeds the deviation threshold, the backtracking analysis finds that the "mathematics subject performance" feature has too high weight in the model, and the statistical data shows that the average score of boys in the subject is generally higher than that of girls, which leads to the overall evaluation of the female group being too low, thereby causing systematic underestimation of the student group that does not occupy the advantage, and therefore the analysis result is marked as having potential deviation.

[0151] Step S403: According to the feature importance distribution, dynamically configure fairness constraint conditions in the large language model, and regenerate the corrected learning profile analysis result based on the fairness constraint conditions.

[0152] The fairness constraint condition refers to the model adjustment rule set based on the feature importance distribution, which is used to limit the influence degree of certain features on the analysis result to reduce the risk of deviation.

[0153] According to the feature importance distribution obtained by backtracking analysis, the weight distribution of the relevant features in the large language model is automatically adjusted, the fairness constraint condition is dynamically introduced in the process of generating the learning situation analysis result, the high-impact bias features are inhibited, and the influence of the features favorable to fairness is enhanced, so as to reduce the influence of the original bias in the corrected learning situation analysis result.

[0154] For example, according to the feature importance distribution, it is found that the "classroom speaking frequency" has too great an influence on the learning situation analysis result, so the weight of this feature in the large language model is dynamically reduced, and the influence weights of "assignment completion quality" and "knowledge point mastery degree" are increased, so as to generate a more fair corrected analysis result.

[0155] Step S404: simultaneously display the learning situation analysis result with potential bias and the corrected learning situation analysis result in the visualization interface.

[0156] Simultaneously displaying the learning situation analysis result with potential bias and the corrected learning situation analysis result in the visualization interface is to clearly compare the differences of the large language model before and after the fairness adjustment, and help teachers or students understand the bias source and correction effect.

[0157] By using the above technical solution, the fairness index detection mechanism is introduced, the feature importance analysis and dynamic constraint configuration are combined, the potential bias in the learning situation analysis result can be identified and corrected, the transparency and explainability of the analysis process are improved, the adaptability and fairness of the system to different student groups are enhanced while ensuring the accuracy of the analysis, and more fair and reliable learning situation support is provided for teaching decision-making.

[0158] The embodiment of the application discloses a fairness and accuracy adaptive balancing method. Figure 5 The method comprises the following steps:

[0159] Step S501: measure the model accuracy change amount when the fairness constraint condition is applied.

[0160] The model accuracy change amount refers to the change value of the model accuracy before and after the introduction of the fairness constraint condition, and the formula is ΔA = (A0-A1) / A0, wherein A0 represents the original model accuracy, and A1 represents the model accuracy under the fairness constraint condition.

[0161] For example, when the fairness constraint condition is not applied, the model accuracy is 90%, and after the fairness constraint condition is applied, the accuracy decreases to 81%, and the model accuracy change amount is 10%.

[0162] Step S502: When the model accuracy change amount decreases by more than the accuracy threshold, start the fairness-accuracy evaluation process, and output the key performance data of the large language model, including the original model accuracy, the model accuracy under fairness constraints, and the prediction difference degree between sensitive attribute groups.

[0163] The fairness-accuracy evaluation process refers to a performance analysis mechanism triggered when the model accuracy decrease amplitude exceeds the accuracy threshold. By comparing the key data such as the original model accuracy, the model accuracy under fairness constraints, and the prediction difference degree between sensitive attribute groups, the performance change of the model after introducing fairness constraints is evaluated to determine the balance between fairness improvement and accuracy maintenance, providing a basis for the generation of subsequent mediation schemes.

[0164] The prediction difference degree between sensitive attribute groups represents the difference between the prediction results of the model for student groups with different sensitive attributes, such as gender and academic foundation, under different model constraint conditions.

[0165] For example, if the model accuracy change amount exceeds the accuracy threshold, the output is that the original model accuracy is 90%, the accuracy after applying fairness constraints is 83%, and the prediction difference degree between sensitive attribute groups decreases from 0.35 to 0.15.

[0166] Step S503: Calculate the relationship between the fairness improvement degree and the model accuracy change amount based on the key performance data, and generate multiple selectable mediation schemes, which contain different combinations of fairness and accuracy.

[0167] The fairness improvement degree is used to measure the improvement effect of the model in fairness, and the formula is ΔF = (D0-D1) / D0, where D0 represents the prediction difference degree of the model for different sensitive attribute groups without applying fairness constraints, and D1 represents the prediction difference degree of the model for different sensitive attribute groups after applying fairness constraints.

[0168] Based on the two key indicators of fairness improvement degree and model accuracy change amount, the fairness improvement degree and model accuracy are used as optimization targets, and a multi-objective optimization algorithm is used to find the optimal solution set between them, generating multiple mediation schemes, each corresponding to a specific combination of fairness improvement degree and model accuracy.

[0169] Exemplarily, according to the detected fairness improvement degree of 40%, the model accuracy decreases by 8%, and on this basis, the constraint parameter is adjusted to generate three mediation schemes: scheme one, the model accuracy decreases by 3% when the fairness improvement degree is 30%, scheme two, the model accuracy decreases by 10% when the fairness improvement degree is 50%, and scheme three, the model accuracy decreases by 15% when the fairness improvement degree is 60%. Each scheme corresponds to a different combination of fairness improvement degree and model accuracy, and the user can select according to actual needs.

[0170] Step S504: According to the bias parameter of the mediation scheme, the updated constraint strength parameter is dynamically calculated, and the constraint strength parameter is fed back to the fairness constraint condition for adaptive adjustment. The bias parameter of 0 represents accuracy priority, and the bias parameter of 1 represents fairness priority.

[0171] The bias parameter refers to the parameter set by the user according to actual needs, which is used for the priority of the model between fairness and accuracy, and the value range is [0, 1].

[0172] The constraint strength parameter refers to the parameter for controlling the strength of the fairness constraint condition in the model. The higher the value, the stronger the constraint. The calculation formula is: λ = λa + (1- λ)(1-a), where λ represents the current constraint strength parameter, and a represents the set bias parameter.

[0173] According to the bias parameter and the current constraint strength parameter selected by the user, the constraint strength parameter is dynamically adjusted by weighted calculation according to the formula. When the bias parameter is 0, it is completely inclined to retain the model accuracy, and when the bias parameter is 1, it is completely inclined to improve the fairness, so as to realize the adaptive mediation of the fairness constraint condition to balance the model performance.

[0174] Exemplarily, the current constraint strength parameter λ = 0.6, the user sets the bias parameter a = 0.8 (biased towards fairness), and the updated constraint strength parameter is obtained by substituting the formula: λ = 0.6x0.8 + (1-0.6)(1-0.8) = 0.56.

[0175] By adopting the above technical scheme, the model performance detection and multi-objective optimization mechanism are introduced, which can dynamically adjust the fairness constraint strength under the premise of ensuring the overall accuracy of the model, realize the adaptive balance of fairness and accuracy, improve the flexibility and applicability of the system in different application scenarios, ensure that the learning situation analysis result maintains high analysis reliability while considering fairness, and enhance the controllability and actual landing value of the model.

[0176] Embodiments of the present application disclose a kind of teacher-student teaching conflict intelligent mediation method. Referring to Figure 6 , the method comprises:

[0177] Step S601: Extract at least one difference feature parameter between the teacher and the student, the difference feature parameter including a knowledge point coverage difference, a training intensity difference, or a cognitive load difference.

[0178] The difference feature parameter is a quantitative index reflecting the understanding deviation or cognitive difference between the teacher's suggestion and the student's feedback.

[0179] The knowledge point coverage difference is obtained by comparing the knowledge point distribution of the teacher's suggestion and the attention point distribution of the student's feedback, the training intensity difference is calculated by quantifying the difference value of the training time of the suggestions of both parties, and the cognitive load difference is determined based on the difference between the evaluation scores of the task difficulty of both parties.

[0180] By natural language processing on the teacher-student exchange text, the knowledge point content, the training time description, and the task difficulty evaluation in the teacher's suggestion and the student's feedback are extracted by using text classification and keyword recognition technology, and the differences in knowledge point coverage, training intensity requirement, and cognitive load of the two are compared respectively, forming feature parameters such as knowledge point coverage difference, training intensity difference, and cognitive load difference, which are used for subsequent teaching conflict analysis and mediation scheme generation.

[0181] Step S602: Calculate a comprehensive conflict index based on the difference feature parameters.

[0182] The comprehensive conflict index is a numerical value that measures the overall conflict degree between the teacher and the student in terms of teaching content, training arrangement, and cognitive understanding, and the formula is: C=W1*D k +W2*D t +W3*D c , where D k is the knowledge point coverage difference, D t is the training intensity deviation value, and D c is the cognitive load difference coefficient, W1, W2, and W3 are weight coefficients corresponding to each difference feature parameter, and satisfy W1+W2+W3=1.

[0183] By multiplying each difference feature parameter by the corresponding weight coefficient and summing the values, the overall conflict degree between the teacher and the student in terms of knowledge point coverage, training intensity, and cognitive load is quantified, and the specific calculation method is to weight sum the knowledge point coverage difference, the training intensity difference, and the cognitive load difference by assigning them with preset weights, to obtain a comprehensive value as the conflict index for judging the severity of the teaching conflict.

[0184] For example, if D k =0.4, D t =1, and D c=2, and the weights are set as W1=0.5, W2=0.3, and W3=0.2, then the comprehensive conflict index is C=0.5*0.4+0.3*1+0.2*2=0.2+0.3+0.4=0.9.

[0185] Step S603: If the comprehensive conflict index exceeds the comprehensive threshold, based on the learning situation analysis result, the first expected effect of the teacher's suggestion and the second expected effect of the student feedback scheme are respectively predicted, and a coordination scheme is generated.

[0186] The first expected effect refers to the teaching effectiveness that can be achieved in terms of knowledge mastery improvement and learning goal achievement based on the teaching suggestion proposed by the teacher and the learning situation analysis result.

[0187] The second expected effect refers to the learning effectiveness that can be achieved in terms of learning adaptability and task completion degree based on the learning scheme fed back by the student and the learning situation analysis result.

[0188] By using a multi-objective optimization algorithm to find a compromise solution among multiple conflicting objectives, the knowledge point coverage, training intensity of the teacher's suggestion, and the cognitive load and acceptance degree of the student feedback scheme are quantitatively modeled, and then under the constraint conditions of meeting the actual learning situation and teaching requirements, a plurality of feasible solutions are calculated by iterative calculation, and the optimal compromise solution that takes into account the teaching effectiveness and student adaptability is selected as the coordination suggestion.

[0189] Step S604: According to the coordination scheme, the display content and form of the coordination scheme are dynamically configured, and the learning situation analysis result is optimized.

[0190] According to the identity of the teacher and the student and the type of the terminal, the coordination scheme is presented in different ways to ensure the pertinence and effectiveness of information transmission. The teaching strategy optimization suggestion including the class overall knowledge mastery degree improvement curve is pushed to the teacher terminal, and the personalized adjustment scheme integrating personal learning characteristic data is output to the student terminal.

[0191] For example, the teacher is shown a curve graph of the class overall knowledge mastery degree from 60% to 70%, and the student is pushed a personalized suggestion of "focusing on reviewing application problems this week, and training 1.5 hours per day".

[0192] By using the above technical solutions, the difference characteristics such as knowledge point coverage, training intensity, and cognitive load between the teacher and the student are extracted, the coordination scheme is generated, and the display content and form are dynamically configured, which can effectively identify and alleviate the deviation in learning situation cognition between the teacher and the student, improve the pertinence and adaptability of teaching intervention, and enhance the efficiency of teacher-student interaction and the intelligent level of teaching decision-making.

[0193] The embodiment of the present application discloses a teaching mediation failure traceability intervention method. Referring to Figure 7The method comprises:

[0194] Step S701: When it is detected that the continuous N coordination schemes do not achieve the expected effect, triggering the traceability process.

[0195] The traceability process refers to an analysis mechanism that is automatically started when the system detects that the continuous N coordination schemes do not achieve the expected effect.

[0196] For example, if the coordination schemes generated by the system for three consecutive times fail to improve the student's knowledge mastery to the target value, the traceability process is triggered.

[0197] Step S702: Based on the traceability process, extract the historical feature parameters in the coordination scheme, and determine the feature parameters of the dominant conflict.

[0198] The historical feature parameters refer to the teacher-student teaching conflict related parameters extracted in the previous coordination process, such as knowledge point coverage difference, training intensity difference, cognitive load difference, etc.

[0199] The feature parameters of the dominant conflict refer to the feature parameters that have the greatest impact on the conflict in multiple coordination failures.

[0200] By analyzing the historical feature parameters in the previous coordination scheme through attention weight distribution, the influence degree of parameters such as knowledge point coverage difference, training intensity difference, and cognitive load difference in the coordination failure process is quantitatively sorted. The feature parameters with higher attention weight indicate that they play a leading role in the coordination conflict, thereby identifying the core conflict factors leading to coordination failure.

[0201] For example, the knowledge point coverage difference, training intensity difference, and cognitive load difference in the last three coordinations are extracted, and it is found through attention weight distribution that the weight of the knowledge point coverage difference is the highest, indicating that the knowledge point coverage difference is the main reason for the current coordination failure.

[0202] Step S703: Generating a conflict traceability report according to the feature parameters, the conflict traceability report including the historical change curve of the feature parameters of the dominant conflict and the matching degree analysis of the related teaching resources.

[0203] The conflict traceability report refers to an analysis report used to record the reasons for coordination failure, display the change trend of the feature parameters of the dominant conflict, and evaluate the adaptation of teaching resources.

[0204] The historical change curve of the feature parameters of the dominant conflict refers to displaying the change of the feature parameters of the dominant conflict in the previous coordination in the form of a curve to help analyze the evolution trend.

[0205] The matching degree analysis of the related teaching resources refers to evaluating the matching degree between the current teaching resources, such as courseware, exercises, teaching methods, and student learning needs.

[0206] For example, the generated conflict trace report indicates that the "knowledge point coverage difference" is continuously high, and plots its change curve in three mediations, while analyzing the insufficient explanation of the related knowledge points in the teaching resources, with a matching degree of only 60%.

[0207] Step S704: Push an alarm signal and a conflict trace report to the administrator terminal, and lock the automatic coordination function on the student terminal and the teacher terminal.

[0208] The alarm signal refers to the prompt information sent by the system to the administrator terminal after detecting continuous failure of coordination, prompting manual intervention.

[0209] The automatic coordination function refers to the function of automatically pushing coordination schemes to teachers and students, and locking the automatic coordination function can prevent the continuous pushing of invalid schemes.

[0210] For example, the system sends a prompt to the administrator terminal: "coordination failure times exceed the standard, please check the trace report and intervene", and suspends the pushing of coordination schemes to the teacher terminal and the student terminal, preventing the continuous pushing of invalid schemes.

[0211] Step S705: In response to receiving a manual confirmation instruction, unlock the automatic coordination function on the student terminal and the teacher terminal.

[0212] The manual confirmation instruction refers to the instruction sent by the administrator to the system after checking the trace report and confirming the processing opinion, which is used to restore the automatic sending of coordination scheme function.

[0213] For example, the administrator clicks the "confirm processing complete" button, and after the system receives the manual confirmation instruction, it restores the pushing of coordination schemes to the teacher terminal and the student terminal, and the system regains the ability to push coordination schemes.

[0214] By adopting the above technical scheme, the continuous monitoring and trace mechanism of coordination effect is introduced, which can automatically trigger trace analysis when multiple coordination is invalid, accurately identify the key factors leading to conflict, and generate a trace report. Combined with manual intervention, it improves the reliability and safety of system decision-making, effectively prevents the repeated execution of invalid mediation, and enhances the intelligent management level and abnormal processing capability of the teaching intervention system.

[0215] Based on the same inventive concept, the embodiments of the present application provide a learning situation analysis system based on a large language model, please refer to Figure 8 The system comprises:

[0216] The acquisition module 801 is configured to acquire learning data and interaction information.

[0217] The memory 802 is configured to store the program of the learning situation analysis method based on the large language model.

[0218] The processor 803 can load and execute the program in the memory to implement the learning situation analysis method based on a large language model.

[0219] By adopting the technical solution, the acquisition module collects student learning data in real time, the processor efficiently executes the learning situation analysis algorithm based on a large language model, and the memory continuously optimizes the analysis model and the data feature library, realizing intelligent processing of the whole process from data collection to learning situation diagnosis to teaching intervention, significantly improving the analysis accuracy and pertinence, and significantly enhancing the teaching decision efficiency, providing efficient and reliable technical support for precise teaching and individualized teaching in the intelligent education scene.

[0220] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0221] The embodiment of the present application provides a computer readable storage medium, which stores a computer program capable of being loaded and executed by a processor to execute a learning situation analysis method based on a large language model.

[0222] The computer storage medium includes, for example, a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0223] Based on the same inventive concept, the embodiment of the present application provides an intelligent terminal, which includes a memory and a processor, and the memory stores a computer program capable of being loaded and executed by the processor to execute a learning situation analysis method based on a large language model.

[0224] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0225] The above are only preferred embodiments of the present application, not intended to limit the protection scope of the present application, any one feature disclosed in the specification (including the abstract and the drawings) can be replaced by other equivalent or similar features, unless specifically stated otherwise. That is, unless specifically stated, each feature is only an example of a series of equivalent or similar features.

Claims

1. A learning situation analysis method based on a large language model, characterized by, The method comprises the following steps: Collecting learning data of students, the learning data comprising structured learning data and unstructured learning data, the structured learning data comprising test scores and homework completion, and the unstructured learning data comprising classroom behavior data, learning process data, student self-evaluation and peer evaluation data, and teacher teaching records; Preprocessing the learning data to generate standardized data for analysis; Receiving input interaction information to generate analysis requirement parameters, the interaction information comprising analysis requirements of teachers or feedback information of students; Based on a large language model, performing semantic understanding and deep analysis on the standardized data and the analysis requirement parameters to obtain learning situation analysis results; Displaying the learning situation analysis results in a visual manner; Detecting fairness indicators of the learning situation analysis results, the fairness indicators comprising at least one of score distribution difference of different gender student groups, similarity of suggestions of different academic foundation student groups, and attention deviation value of special needs student groups, the fairness indicators being quantitative indicators for measuring whether the learning situation analysis results have systematic bias among different student groups; If any of the fairness indicators exceeds a bias threshold, generating a feature importance distribution according to the fairness indicators, and recording the learning situation analysis results as learning situation analysis results with potential bias, the feature importance distribution containing influence scores of input features of the large language model on final analysis results; According to the feature importance distribution, dynamically configuring fairness constraint conditions in the large language model, and regenerating corrected learning situation analysis results based on the fairness constraint conditions, the fairness constraint conditions being model adjustment rules set based on the feature importance distribution, used to limit the influence of certain features on learning situation analysis results to reduce bias risk; Simultaneously displaying the learning situation analysis results with potential bias and the corrected learning situation analysis results in a visual interface; Measuring a model accuracy change amount when the fairness constraint conditions are applied, the model accuracy change amount being a change value of model accuracy before and after the fairness constraint conditions are introduced, formula: ΔA = (A0-A1) / A0, wherein A0 represents the original model accuracy, and A1 represents the model accuracy under the fairness constraint conditions; When the model accuracy change amount decreases by more than an accuracy threshold, starting a fairness-accuracy evaluation process, and outputting key performance data of the large language model, the key performance data comprising the original model accuracy, the model accuracy under the fairness constraint conditions, and prediction difference between sensitive attribute groups. According to the relationship of the key performance data and the fairness improvement degree, a plurality of optional mediation schemes are generated, the mediation schemes including different combinations of fairness and accuracy, and the fairness improvement degree is used to measure the improvement effect of the model in fairness, and the formula is ΔF = (D0-D1) / D0, wherein D0 represents the prediction difference of the model for different sensitive attribute groups without applying the fairness constraint condition, and D1 represents the prediction difference of the model for the different sensitive attribute groups after applying the fairness constraint condition; According to the mediation scheme, a bias parameter is selected, and an updated constraint strength parameter is dynamically calculated, and the constraint strength parameter is fed back to the fairness constraint condition for adaptive adjustment, the bias parameter is 0, which represents that the accuracy is preferred, the bias parameter is 1, which represents that the fairness is preferred, and the constraint strength parameter refers to a parameter for adjusting the strength of the fairness constraint condition in the model, and the larger the parameter value is, the stronger the fairness constraint condition applied to the model is, and the formula is: λ = λa + (1-λ)(1-a), wherein λ represents the current constraint strength parameter, and a represents the set bias parameter. 2.The learning situation analysis method based on a large language model of claim 1, wherein, After the unstructured learning data is preprocessed, the method further includes: detecting the consistency of the classroom behavior data and the learning process data; if the consistency does not meet the preset standard, determining that there is a logical conflict when the classroom behavior data is higher than a first data threshold and the learning process data is lower than a second data threshold; in response to the logical conflict, analyzing effective participation features in the monitoring video data in the conflict time period to obtain an effectiveness score, the effective participation features including an angle between a face orientation and a blackboard or a teacher, a relevance between a body movement and a learning task, and an effective speech in group interaction; if the effectiveness score exceeds an effectiveness threshold, increasing a first weight coefficient of the classroom behavior data and decreasing a second weight coefficient of the learning process data; if the effectiveness score does not exceed the effectiveness threshold, decreasing a third weight coefficient of the classroom behavior data and increasing a fourth weight coefficient of the learning process data; weighting and fusing the classroom behavior data and the learning process data according to the first weight coefficient, the second weight coefficient, the third weight coefficient and the fourth weight coefficient to generate the standardized data. 3.The learning situation analysis method based on a large language model of claim 2, wherein, The detection of the abnormal situation of the classroom behavior data includes: collecting state features of students, the state features including body movement amplitude features, face orientation angle change features and group interaction state features; when the state features exceed a safety threshold, determining that there is an abnormal teaching event; if there is the abnormal teaching event, freezing the classroom behavior data, marking the classroom behavior data as invalid data segment and removing the invalid data segment; taking valid data before the generation time of the invalid data segment as regular data; based on the regular data, generating and calculating missing data and a confidence; if the confidence exceeds a threshold, adding the missing data to the classroom behavior data; if the confidence does not exceed a threshold, freezing the missing data. 4.The learning situation analysis method based on a large language model of claim 1, wherein, The method further comprises: extracting at least one difference feature parameter between the teacher and the student, the difference feature parameter including a knowledge point coverage difference, a training intensity difference, or a cognitive load difference; based on the difference feature parameter, calculating a comprehensive conflict index; if the comprehensive conflict index exceeds a comprehensive threshold, based on the learning situation analysis result, predicting a first expected effect of the teacher's suggestion and a second expected effect of the student's feedback scheme respectively, and generating a coordination scheme; according to the coordination scheme, dynamically configuring the display content and form of the coordination scheme, and optimizing the learning situation analysis result. 5.The learning situation analysis method based on a large language model according to claim 4, characterized in that, when the generation of the coordination scheme fails, comprising: when it is detected that the coordination scheme does not achieve the expected effect for N consecutive times, triggering a traceability process; based on the traceability process, extracting historical feature parameters in the coordination scheme to determine the feature parameters of the dominant conflict; generating a conflict traceability report according to the feature parameters, the conflict traceability report including a duration change curve of key conflict parameters and a matching degree analysis of related teaching resources; pushing an alarm signal and the conflict traceability report to an administrator terminal, and locking the automatic coordination function on the student terminal and the teacher terminal; in response to receiving a manual confirmation instruction, unlocking the automatic coordination function on the student terminal and the teacher terminal. 6.A learning situation analysis system based on a large language model, characterized by The system is used to execute the learning situation analysis method based on the large language model as claimed in any one of claims 1 to 5, comprising: an acquisition module for acquiring learning data and interaction information; a memory for storing the program of the learning situation analysis method based on the large language model; a processor, the program in the memory can be loaded and executed by the processor and realize the learning situation analysis method based on the large language model.

7. A smart terminal, characterized in that including a memory and a processor, the memory has stored a computer program which can be loaded and executed by the processor to execute the method as claimed in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, a computer program which can be loaded and executed by the processor to execute the method as claimed in any one of claims 1 to 5.

Images