Student multi-dimensional data management method and system based on cross-model semantic fusion

Abstract

The application discloses a student multi-dimensional data management method and system based on cross-model semantic fusion, and the method comprises the following steps: obtaining service record data of multiple dimensions generated by a student object in multiple education service scenes; inputting each service record data into an education semantic analysis model of a corresponding education service scene to obtain a corresponding record analysis result; determining a semantic matching rule between any two service record data according to model matching between multiple education semantic analysis models; and performing fusion processing on all record analysis results according to the semantic matching rule to obtain a multi-dimensional data analysis result corresponding to the student object. It can be seen that the application can realize semantic association and comprehensive evaluation of cross-scene service records, improve the comprehensiveness and accuracy of student multi-dimensional data performance analysis, and reduce the risk of data analysis and evaluation deviation caused by scene isolation.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for managing student multidimensional data based on cross-model semantic fusion. Background Technology

[0002] With the rapid growth in demand for digital education and personalized assessment, educational institutions and enterprises are increasingly emphasizing data recording of students' multidimensional performance across various service scenarios. A key technical issue is how to efficiently and effectively manage this recorded data. Existing technologies typically use independent semantic analysis models or manual annotation to process data from different scenarios, generating evaluation reports based on simple summaries or dimensional splicing to support student performance evaluation. However, existing solutions lack a unified semantic analysis framework for data recorded across different educational service scenarios, dynamic determination of matching rules between models, and a deep fusion mechanism for multi-model analysis results. This makes it difficult to achieve semantic association and multidimensional comprehensive evaluation of cross-scenario service records. Commonly used isolated or shallow splicing methods fail to effectively capture implicit relationships and semantic consistency between scenarios, resulting in insufficient comprehensiveness and accuracy in student multidimensional performance analysis. Isolated scenarios can easily lead to evaluation biases or omissions of important information, limiting the scientific rigor, objectivity, and personalized guidance value of educational evaluation systems. Therefore, existing technologies have shortcomings that urgently need to be addressed. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a student multidimensional data management method and system based on cross-model semantic fusion, which can realize semantic association and comprehensive evaluation of cross-scenario service records, improve the comprehensiveness and accuracy of student multidimensional data performance analysis, and reduce the risk of data analysis and evaluation bias caused by scenario isolation.

[0004] To address the aforementioned technical problems, the first aspect of this invention discloses a method for managing multidimensional student data based on cross-model semantic fusion, the method comprising: Acquire service record data from multiple dimensions generated by student objects across various educational service scenarios; Each service record data is input into the corresponding educational semantic analysis model of the educational service scenario to obtain the corresponding record analysis results; Based on the model matching between multiple educational semantic analysis models, determine the semantic matching rules between any two service record data; The analysis results of all the records are fused according to the semantic matching rules to obtain the multidimensional data analysis results corresponding to the student object.

[0005] As an optional implementation, in a first aspect of the invention, the service record data includes at least one of visual data, sound data, tactile data, light reflection data, text data, and computer operation data.

[0006] As an optional implementation, in the first aspect of the present invention, the educational service scenario is a classroom service scenario, a self-study service scenario, an exercise service scenario, an examination service scenario, or an extracurricular activity service scenario.

[0007] As an optional implementation, in the first aspect of the present invention, the step of inputting each of the service record data into the educational semantic analysis model of the corresponding educational service scenario to obtain the corresponding record analysis result includes: For each service record data, determine the educational semantic analysis model corresponding to the educational service scenario for that service record data; the educational semantic analysis model includes a data vectorization model and a semantic analysis model. The service record data is input into the data vectorization model to obtain the output semantic relevance vector; The semantic relevance vectors are input into the semantic analysis model to obtain the record analysis results corresponding to the service record data; the record analysis results include quantitative data corresponding to multiple evaluation dimensions of the corresponding education service scenario; the semantic analysis model is trained using a training dataset that includes multiple training service record data in the corresponding education service scenario, as well as corresponding semantic relevance vector annotations and multi-dimensional analysis annotations.

[0008] As an optional implementation, in a first aspect of the invention, the data vectorization model includes a data semantic prediction network and a vectorized search network; the data semantic prediction network is used to predict the semantic content corresponding to the service record data; the vectorized search network is used to perform the following steps: Calculate the first similarity between the historical processing data and the service record data corresponding to each candidate vectorization model; the candidate vectorization model is trained using a training dataset that includes training record data from multiple educational service scenarios, as well as corresponding semantic content annotations and semantically related vector annotations; The semantic content and the preset template record data are input into each of the candidate vectorization models to obtain the corresponding semantic output vector; Calculate the second similarity between the historical output semantic vector and the semantic output vector corresponding to each candidate vectorized model; Calculate the product of the first similarity and the second similarity to obtain the model matching degree of each candidate vectorized model; The candidate vectorized model with the highest model matching degree is determined as the target vectorized model; The semantic content and the service record data are input into the target vectorization model to obtain the output semantically related vector.

[0009] As an optional implementation, in the first aspect of the present invention, determining the semantic matching rule between any two service record data based on model matching among multiple educational semantic analysis models includes: For any two service record data, multiple preset standard student record data are input into the educational semantic analysis model corresponding to the two service record data to obtain the analysis results of the two educational semantic analysis models for each standard student record data; For each of the standard student record data, calculate the similarity of the results for each evaluation dimension between the analysis results of the two educational semantic analysis models for the standard student record data; The rating dimensions with similarity scores higher than a preset similarity threshold are selected to obtain the set of associated dimensions corresponding to the standard student record data; Calculate the intersection of the associated dimension sets of all the calibration student record data to obtain the evaluation dimension set corresponding to the two service record data; All evaluation dimensions in the evaluation dimension set corresponding to the two service record data are determined as association dimensions, and the semantic matching rules corresponding to the two service record data are obtained.

[0010] As an optional implementation, in the first aspect of the present invention, the step of fusing all the record analysis results according to the semantic matching rules to obtain the multidimensional data analysis results corresponding to the student object includes: For any two record analysis results, cross-dimensional correction is performed on the two record analysis results based on the semantic matching rules to obtain the corrected dimension result; Record all the results of the correction dimensions and the process of the cross-dimensional correction, and determine them as the analysis data corresponding to the student object; The analysis data is input into the trained LLM model to obtain the performance evaluation of the student object in multiple dimensions; the LLM model is trained using a training dataset that includes multiple training analysis data and corresponding performance evaluation labels.

[0011] As an optional implementation, in the first aspect of the invention, the step of performing cross-dimensional correction on the two record analysis results based on the semantic matching rules to obtain the corrected dimensional result includes: For each evaluation dimension, based on the semantic matching rules between the service record data corresponding to the two record analysis results, it is determined whether the evaluation dimension is a related dimension, and a first judgment result is obtained; If the first judgment result is yes, calculate the weighted sum of the evaluation data of the two record analysis results in the evaluation dimension to obtain the correction dimension result corresponding to the evaluation dimension; If the first judgment result is negative, calculate the proportion of the number of multiple preset similarity dimensions corresponding to the evaluation dimension that belong to the associated dimensions in the corresponding semantic matching rule; Determine whether the quantity ratio is greater than a preset ratio threshold to obtain a second determination result; If the second judgment result is yes, calculate the weighted sum of the evaluation data of the two recorded analysis results in the evaluation dimension to obtain the correction dimension result corresponding to the evaluation dimension; If the second judgment result is negative, stop the correction step for that evaluation dimension.

[0012] A second aspect of this invention discloses a student multidimensional data management system based on cross-model semantic fusion, the system comprising: The acquisition module is used to acquire multi-dimensional service record data generated by student objects in multiple educational service scenarios; The analysis module is used to input the service record data of each service into the educational semantic analysis model of the corresponding educational service scenario to obtain the corresponding record analysis results. The determination module is used to determine the semantic matching rules between any two service record data based on the model matching between multiple educational semantic analysis models; The fusion module is used to fuse all the record analysis results according to the semantic matching rules to obtain the multidimensional data analysis results corresponding to the student object.

[0013] As an optional implementation, in a second aspect of the invention, the service record data includes at least one of visual data, sound data, tactile data, light reflection data, text data, and computer operation data.

[0014] As an optional implementation, in the second aspect of the present invention, the educational service scenario is a classroom service scenario, a self-study service scenario, an exercise service scenario, an examination service scenario, or an extracurricular activity service scenario.

[0015] As an optional implementation, in the second aspect of the present invention, the specific method by which the analysis module inputs each of the service record data into the corresponding educational semantic analysis model of the educational service scenario to obtain the corresponding record analysis results includes: For each service record data, determine the educational semantic analysis model corresponding to the educational service scenario for that service record data; the educational semantic analysis model includes a data vectorization model and a semantic analysis model. The service record data is input into the data vectorization model to obtain the output semantic relevance vector; The semantic relevance vectors are input into the semantic analysis model to obtain the record analysis results corresponding to the service record data; the record analysis results include quantitative data corresponding to multiple evaluation dimensions of the corresponding education service scenario; the semantic analysis model is trained using a training dataset that includes multiple training service record data in the corresponding education service scenario, as well as corresponding semantic relevance vector annotations and multi-dimensional analysis annotations.

[0016] As an optional implementation, in a second aspect of the invention, the data vectorization model includes a data semantic prediction network and a vectorized search network; the data semantic prediction network is used to predict the semantic content corresponding to the service record data; the vectorized search network is used to perform the following steps: Calculate the first similarity between the historical processing data and the service record data corresponding to each candidate vectorization model; the candidate vectorization model is trained using a training dataset that includes training record data from multiple educational service scenarios, as well as corresponding semantic content annotations and semantically related vector annotations; The semantic content and the preset template record data are input into each of the candidate vectorization models to obtain the corresponding semantic output vector; Calculate the second similarity between the historical output semantic vector and the semantic output vector corresponding to each candidate vectorized model; Calculate the product of the first similarity and the second similarity to obtain the model matching degree of each candidate vectorized model; The candidate vectorized model with the highest model matching degree is determined as the target vectorized model; The semantic content and the service record data are input into the target vectorization model to obtain the output semantically related vector.

[0017] As an optional implementation, in a second aspect of the invention, the determining module determines the specific method of the semantic matching rule between any two service record data based on model matching among multiple educational semantic analysis models, including: For any two service record data, multiple preset standard student record data are input into the educational semantic analysis model corresponding to the two service record data to obtain the analysis results of the two educational semantic analysis models for each standard student record data; For each of the standard student record data, calculate the similarity of the results for each evaluation dimension between the analysis results of the two educational semantic analysis models for the standard student record data; The rating dimensions with similarity scores higher than a preset similarity threshold are selected to obtain the set of associated dimensions corresponding to the standard student record data; Calculate the intersection of the associated dimension sets of all the calibration student record data to obtain the evaluation dimension set corresponding to the two service record data; All evaluation dimensions in the evaluation dimension set corresponding to the two service record data are determined as association dimensions, and the semantic matching rules corresponding to the two service record data are obtained.

[0018] As an optional implementation, in a second aspect of the invention, the specific method by which the fusion module performs fusion processing on all the record analysis results according to the semantic matching rules to obtain the multidimensional data analysis results corresponding to the student object includes: For any two record analysis results, cross-dimensional correction is performed on the two record analysis results based on the semantic matching rules to obtain the corrected dimension result; Record all the results of the correction dimensions and the process of the cross-dimensional correction, and determine them as the analysis data corresponding to the student object; The analysis data is input into the trained LLM model to obtain the performance evaluation of the student object in multiple dimensions; the LLM model is trained using a training dataset that includes multiple training analysis data and corresponding performance evaluation labels.

[0019] As an optional implementation, in a second aspect of the invention, the fusion module performs cross-dimensional correction on the two record analysis results based on the semantic matching rules to obtain the corrected dimensional result in the following specific manner: For each evaluation dimension, based on the semantic matching rules between the service record data corresponding to the two record analysis results, it is determined whether the evaluation dimension is a related dimension, and a first judgment result is obtained; If the first judgment result is yes, calculate the weighted sum of the evaluation data of the two record analysis results in the evaluation dimension to obtain the correction dimension result corresponding to the evaluation dimension; If the first judgment result is negative, calculate the proportion of the number of multiple preset similarity dimensions corresponding to the evaluation dimension that belong to the associated dimensions in the corresponding semantic matching rule; Determine whether the quantity ratio is greater than a preset ratio threshold to obtain a second determination result; If the second judgment result is yes, calculate the weighted sum of the evaluation data of the two recorded analysis results in the evaluation dimension to obtain the correction dimension result corresponding to the evaluation dimension; If the second judgment result is negative, stop the correction step for that evaluation dimension.

[0020] A third aspect of this invention discloses another student multidimensional data management system based on cross-model semantic fusion, the system comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute some or all of the steps in the student multidimensional data management method based on cross-model semantic fusion disclosed in the first aspect of the present invention.

[0021] The fourth aspect of the present invention discloses a computer storage medium storing computer instructions, which, when invoked, are used to execute some or all of the steps in the student multidimensional data management method based on cross-model semantic fusion disclosed in the first aspect of the present invention.

[0022] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: This invention obtains service record data from multiple educational service scenarios for students and inputs it into corresponding educational semantic analysis models to obtain record analysis results. Based on the matching between models, semantic matching rules are determined, and all analysis results are integrated to generate multidimensional data analysis results. This enables the semantic association and comprehensive evaluation of service records across scenarios, improves the comprehensiveness and accuracy of multidimensional data performance analysis of students, and reduces the risk of data analysis and evaluation bias caused by scenario isolation. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a student multidimensional data management method based on cross-model semantic fusion disclosed in an embodiment of the present invention.

[0025] Figure 2 This is a schematic diagram of the structure of a student multidimensional data management system based on cross-model semantic fusion disclosed in an embodiment of the present invention.

[0026] Figure 3This is a schematic diagram of another student multidimensional data management system based on cross-model semantic fusion disclosed in an embodiment of the present invention. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0029] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0030] This invention discloses a method and system for managing multidimensional student data based on cross-model semantic fusion. It acquires service record data from multiple educational service scenarios for students and inputs it into corresponding educational semantic analysis models to obtain record analysis results. Semantic matching rules are determined based on inter-model matching, and all analysis results are fused to generate multidimensional data analysis results. This enables semantic association and comprehensive evaluation of service records across scenarios, improving the comprehensiveness and accuracy of multidimensional student data performance analysis and reducing the risk of data analysis and evaluation bias caused by scenario isolation. Detailed explanations follow.

[0031] Example 1 Please see Figure 1 , Figure 1 This is a flowchart illustrating a student multidimensional data management method based on cross-model semantic fusion disclosed in an embodiment of the present invention. Figure 1The described student multidimensional data management method based on cross-model semantic fusion can be applied to data processing systems / data processing equipment / data processing servers (wherein, the server includes local processing servers or cloud processing servers). Figure 1 As shown, this student multidimensional data management method based on cross-model semantic fusion can include the following operations: 101. Obtain service record data of multiple dimensions generated by student objects in multiple educational service scenarios.

[0032] Optionally, the service record data includes at least one of visual data, sound data, tactile data, light reflection data, text data, and computer operation data.

[0033] Optionally, the educational service scenarios can be classroom service scenarios, self-study service scenarios, exercise service scenarios, examination service scenarios, or extracurricular activity service scenarios.

[0034] Optional educational service scenarios include, but are not limited to, online classes, offline tutoring, homework submission, exam assessment, after-class Q&A, learning community interaction, learning behavior tracking, or parent feedback records.

[0035] Optionally, service record data from multiple dimensions may include, but are not limited to, learning duration data, answer accuracy data, homework completion quality data, classroom participation data, knowledge point mastery data, learning emotion data, learning path data, or learning resource usage data.

[0036] 102. Input the data of each service record into the corresponding educational semantic analysis model of the educational service scenario to obtain the corresponding record analysis results.

[0037] Optionally, the educational semantic analysis model can be configured differently according to the specific service scenario type. For example, the online classroom scenario uses a video + text multimodal analysis model, and the homework submission scenario uses a text understanding + rating regression model. This invention does not impose any limitations.

[0038] 103. Based on model matching between multiple educational semantic analysis models, determine the semantic matching rules between any two service record data.

[0039] Optionally, model matching can be evaluated based on model structure similarity, parameter distribution similarity, or output consistency on the same input, and this invention does not limit the scope of the evaluation.

[0040] Optionally, semantic matching rules may include a shared set of evaluation dimensions, a dimension weight mapping table, cross-dimensional correction coefficients, or a matrix of inter-dimensional dependencies; this invention does not impose any limitations on these rules.

[0041] 104. Based on semantic matching rules, the analysis results of all records are fused to obtain the multidimensional data analysis results corresponding to the student objects.

[0042] Optionally, the fusion process may include dimension alignment, weighted fusion, conflict resolution, or temporal smoothing, and this invention does not limit it.

[0043] Optionally, the results of multidimensional data analysis can be further standardized to output evaluation vectors or multidimensional scoring matrices with unified dimensions, supporting subsequent visualization or trend analysis. This invention does not impose any limitations on this.

[0044] As can be seen, the above-mentioned embodiments of the invention obtain service record data of students in multiple educational service scenarios and input them into the corresponding educational semantic analysis model to obtain record analysis results. Based on the matching between models, semantic matching rules are determined and all analysis results are integrated to generate multidimensional data analysis results. This enables the semantic association and comprehensive evaluation of cross-scenario service records, improves the comprehensiveness and accuracy of students' multidimensional data performance analysis, and reduces the risk of data analysis and evaluation bias caused by scenario isolation.

[0045] As an optional embodiment, the step described above, inputting each service record data into the corresponding educational semantic analysis model for the educational service scenario to obtain the corresponding record analysis results, includes: For each service record data, determine the educational semantic analysis model corresponding to the educational service scenario for that service record data; the educational semantic analysis model includes a data vectorization model and a semantic analysis model. The service record data is input into the data vectorization model to obtain the semantically relevant output vector; The semantic relevance vector is input into the semantic analysis model to obtain the record analysis results corresponding to the service record data.

[0046] Optionally, the recorded analysis results may include quantitative data corresponding to multiple evaluation dimensions for the relevant educational service scenarios.

[0047] Optionally, the semantic analysis model is trained using a training dataset that includes multiple training service record data from the corresponding educational service scenario, as well as corresponding semantically related vector annotations and multi-dimensional analysis annotations.

[0048] Optionally, the training dataset can contain more than 300,000 labeled samples, supporting scene-adaptive fine-tuning; this invention does not impose any limitations.

[0049] Optionally, the process of determining the educational service scenario corresponding to the service record data can be achieved by first identifying the scenario to which the service record belongs using a scenario classifier, and then loading the corresponding model. This invention does not impose any limitations on this process.

[0050] Optionally, the data vectorization model can be a multimodal Transformer encoder that supports multiple input modalities such as text, numerical values, and time series data; however, this invention does not impose any limitations on this model.

[0051] Optionally, multiple evaluation dimensions may include knowledge mastery, learning initiative, depth of understanding, expression ability, focus, or innovation ability; this invention does not impose any limitations.

[0052] As can be seen, through the above optional embodiments, by determining the educational semantic analysis model corresponding to the service record data scenario, the data is input into the data vectorization model to obtain semantic related vectors, and then input into the semantic analysis model to output multi-dimensional quantitative record analysis results. This achieves accurate semantic parsing and quantification based on scenario-specific models, improves the relevance and professionalism of record analysis results, and reduces the risk of dimensional evaluation distortion caused by the generality of the model.

[0053] As an optional embodiment, in the above steps, the data vectorization model includes a data semantic prediction network and a vectorization search network; the data semantic prediction network is used to predict the semantic content corresponding to the service record data; the vectorization search network is used to perform the following steps: Calculate the first similarity between the historical processing data and service record data corresponding to each candidate vectorization model; optionally, the candidate vectorization model is trained using a training dataset that includes training record data from multiple educational service scenarios, as well as corresponding semantic content annotations and semantically related vector annotations. The semantic content and the preset template record data are input into each candidate vectorization model to obtain the corresponding semantic output vector; Calculate the second similarity between the historical output semantic vector and the semantic output vector corresponding to each candidate vectorized model; Calculate the product of the first similarity and the second similarity to obtain the model matching degree of each candidate vectorized model; The candidate vectorized model with the highest model matching degree is determined as the target vectorized model; The semantic content and service record data are input into the target vectorization model to obtain the output semantically related vectors.

[0054] Optionally, the data semantic prediction network can be a BERT-based sequence labeling model or a T5 generative model; this invention does not limit the specific model.

[0055] Optionally, the first similarity can be the cosine similarity of sentence vectors or the Euclidean distance of feature vectors; this invention does not limit this.

[0056] Optionally, the data recorded in this template can be a standard learning record template, and this invention does not limit it.

[0057] As can be seen, through the above optional embodiments, semantic content is predicted by the data semantic prediction network and the model matching degree of the candidate vectorized models is calculated by the vectorized search network to select the target vectorized model output semantically related vectors. This achieves dynamic vectorization optimization based on semantic prediction and historical matching, improves the expression accuracy and adaptability of semantically related vectors, and reduces the risk of vector distortion caused by fixed vectorized models.

[0058] Optionally, the second similarity can be vector cosine similarity, which is not limited in this invention.

[0059] Optionally, the semantic relevance vector can be a 768-dimensional or 1024-dimensional dense vector, and this invention does not limit it.

[0060] As an optional embodiment, the step above, determining the semantic matching rules between any two service record data based on model matching among multiple educational semantic analysis models, includes: For any two service record data, multiple preset standard student record data are input into the educational semantic analysis model corresponding to the two service record data to obtain the analysis results of the two educational semantic analysis models for each standard student record data. For each standard student record data, calculate the similarity of the results of each evaluation dimension between the analysis results of the two educational semantic analysis models for that standard student record data. The rating dimensions with similarity scores higher than the preset similarity threshold are selected to obtain the set of associated dimensions corresponding to the student record data of this standard. Calculate the intersection of the sets of associated dimensions of all calibration student record data to obtain the set of evaluation dimensions corresponding to the two service record data; All evaluation dimensions in the evaluation dimension set corresponding to the two service record data are identified as association dimensions, and the semantic matching rules corresponding to the two service record data are obtained.

[0061] Optionally, the standard student record data can be a manually constructed standardized learning case library, which is not limited in this invention.

[0062] Optionally, the similarity result can be the reciprocal or cosine similarity of the numerical difference, which is not limited in this invention.

[0063] Optionally, the cross-dimensional correction can be a weighted average, bias compensation, or consistency adjustment, and this invention does not limit it.

[0064] As can be seen, through the above optional embodiments, by inputting standard student record data into the educational semantic analysis model corresponding to two service records, calculating the similarity of the evaluation dimension results, filtering the set of related dimensions, and taking the intersection to obtain semantic matching rules, the automatic extraction of cross-model semantic association rules based on standard samples is realized, improving the objectivity and consistency of the rules, and reducing the risk of matching deviation caused by the subjectivity of manual rules.

[0065] As an optional embodiment, the above steps, including fusing all record analysis results according to semantic matching rules to obtain multidimensional data analysis results corresponding to student objects, include: For any two record analysis results, cross-dimensional correction is performed on the two record analysis results based on semantic matching rules to obtain the corrected dimension results; Record all the results of the correction dimensions and the process of cross-dimensional correction, and identify them as the analysis data corresponding to the student subjects; The analysis data is input into the trained LLM model to obtain the performance evaluation of the student object in multiple dimensions; the LLM model is trained on a training dataset that includes multiple training analysis data and corresponding performance evaluation labels.

[0066] Optionally, the process record may include correction weights, similarity values, and judgment paths, which are not limited in this invention.

[0067] Optionally, the LLM model can be a Chinese-optimized Qwen-72B model, with fine-tuning instructions on 300,000 analysis data-evaluation label pairs; this invention does not impose any limitations.

[0068] As can be seen, through the above optional embodiments, by performing cross-dimensional correction on the record analysis results based on semantic matching rules to obtain the corrected dimension results, and inputting them into the LLM model to output multi-dimensional performance evaluation, a precise comprehensive evaluation of student performance is achieved by integrating the correlation dimension correction with the large model, thereby improving the interpretability and comprehensiveness of the evaluation results and reducing the risk of biased performance evaluation caused by uncorrected dimensions.

[0069] As an optional embodiment, the step above, performing cross-dimensional correction on the two record analysis results based on semantic matching rules to obtain the corrected dimensional result, includes: For each evaluation dimension, based on the semantic matching rules between the service record data corresponding to the analysis results of the two records, it is determined whether the evaluation dimension is a related dimension, and the first judgment result is obtained; If the first judgment result is yes, calculate the weighted sum of the evaluation data of the two record analysis results in that evaluation dimension to obtain the correction dimension result corresponding to that evaluation dimension; If the first judgment result is negative, calculate the proportion of the number of pre-set similarity dimensions corresponding to the evaluation dimension that belong to the related dimensions in the corresponding semantic matching rules; Determine whether the quantity ratio is greater than a preset ratio threshold to obtain a second determination result; If the second judgment result is yes, calculate the weighted sum of the evaluation data of the two record analysis results in that evaluation dimension to obtain the correction dimension result corresponding to that evaluation dimension; If the second judgment result is negative, stop the correction step for that evaluation dimension.

[0070] Optionally, the determination of whether a dimension is related can be made directly by querying the set of related dimensions using computer code, and this invention does not impose any limitations on this.

[0071] Optionally, the weights of the weighted summation can be determined based on the amount of data or the confidence level, and this invention does not impose any limitations.

[0072] As can be seen, through the above optional embodiments, by determining whether the evaluation dimension is a related dimension or the proportion of similar dimensions exceeds the threshold, a weighted summation correction is performed. This achieves accurate dimension correction based on both rules and proportions, improves the pertinence and robustness of the correction process, and reduces the risk of over- or under-correction caused by misjudgment of dimension association.

[0073] Example 2 Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of a student multidimensional data management system based on cross-model semantic fusion, as disclosed in an embodiment of the present invention. Figure 2 The described student multidimensional data management system based on cross-model semantic fusion can be applied to data processing systems / data processing equipment / data processing servers (wherein, the server includes local processing servers or cloud processing servers). For example... Figure 2 As shown, the student multidimensional data management system based on cross-model semantic fusion may include: The acquisition module 201 is used to acquire service record data of multiple dimensions generated by student objects in multiple educational service scenarios.

[0074] The analysis module 202 is used to input the data of each service record into the educational semantic analysis model of the corresponding educational service scenario in order to obtain the corresponding record analysis results.

[0075] The determination module 203 is used to determine the semantic matching rules between any two service record data based on the model matching between multiple educational semantic analysis models.

[0076] The fusion module 204 is used to fuse all record analysis results according to semantic matching rules to obtain multidimensional data analysis results corresponding to student objects.

[0077] As can be seen, the above-mentioned embodiments of the invention obtain service record data of students in multiple educational service scenarios and input them into the corresponding educational semantic analysis model to obtain record analysis results. Based on the matching between models, semantic matching rules are determined and all analysis results are integrated to generate multidimensional data analysis results. This enables the semantic association and comprehensive evaluation of cross-scenario service records, improves the comprehensiveness and accuracy of students' multidimensional data performance analysis, and reduces the risk of data analysis and evaluation bias caused by scenario isolation.

[0078] As an optional embodiment, the service record data includes at least one of visual data, sound data, tactile data, light reflection data, text data, and computer operation data.

[0079] As can be seen, the above optional embodiments limit the details of the service record data to comprehensively characterize the behavioral characteristics of students in the education service scenario, so as to facilitate subsequent semantic analysis and data analysis, help improve the comprehensiveness and accuracy of the multidimensional data performance analysis of students, and reduce the risk of evaluation bias caused by scenario isolation.

[0080] As an optional embodiment, the education service scenario can be a classroom service scenario, a self-study service scenario, an exercise service scenario, an examination service scenario, or an extracurricular activity service scenario.

[0081] As can be seen, the above optional embodiments define the types of educational service scenarios to comprehensively characterize the behavioral features of students in multiple educational service scenarios, so as to facilitate subsequent semantic analysis and data analysis, help improve the comprehensiveness and accuracy of students' multidimensional data performance analysis, and reduce the risk of evaluation bias caused by scenario isolation.

[0082] As an optional embodiment, the analysis module inputs the data of each service record into the educational semantic analysis model of the corresponding educational service scenario to obtain the corresponding record analysis results in the following specific ways: For each service record data, determine the educational semantic analysis model corresponding to the educational service scenario for that service record data; optionally, the educational semantic analysis model includes a data vectorization model and a semantic analysis model. The service record data is input into the data vectorization model to obtain the semantically relevant output vector; The semantic relevance vectors are input into the semantic analysis model to obtain the record analysis results corresponding to the service record data; optionally, the record analysis results include quantitative data corresponding to multiple evaluation dimensions of the corresponding education service scenario; the semantic analysis model is trained using a training dataset that includes multiple training service record data in the corresponding education service scenario, as well as corresponding semantic relevance vector annotations and multi-dimensional analysis annotations.

[0083] As can be seen, through the above optional embodiments, by determining the educational semantic analysis model corresponding to the service record data scenario, the data is input into the data vectorization model to obtain semantic related vectors, and then input into the semantic analysis model to output multi-dimensional quantitative record analysis results. This achieves accurate semantic parsing and quantification based on scenario-specific models, improves the relevance and professionalism of record analysis results, and reduces the risk of dimensional evaluation distortion caused by the generality of the model.

[0084] As an optional embodiment, the data vectorization model includes a data semantic prediction network and a vectorized search network; the data semantic prediction network is used to predict the semantic content corresponding to the service record data; the vectorized search network is used to perform the following steps: Calculate the first similarity between the historical processing data and service record data corresponding to each candidate vectorization model; optionally, the candidate vectorization model is trained using a training dataset that includes training record data from multiple educational service scenarios, as well as corresponding semantic content annotations and semantically related vector annotations. The semantic content and the preset template record data are input into each candidate vectorization model to obtain the corresponding semantic output vector; Calculate the second similarity between the historical output semantic vector and the semantic output vector corresponding to each candidate vectorized model; Calculate the product of the first similarity and the second similarity to obtain the model matching degree of each candidate vectorized model; The candidate vectorized model with the highest model matching degree is determined as the target vectorized model; The semantic content and service record data are input into the target vectorization model to obtain the output semantically related vectors.

[0085] As can be seen, through the above optional embodiments, semantic content is predicted by the data semantic prediction network and the model matching degree of the candidate vectorized models is calculated by the vectorized search network to select the target vectorized model output semantically related vectors. This achieves dynamic vectorization optimization based on semantic prediction and historical matching, improves the expression accuracy and adaptability of semantically related vectors, and reduces the risk of vector distortion caused by fixed vectorized models.

[0086] As an optional embodiment, the determining module determines the specific method of semantic matching rules between any two service record data based on model matching between multiple educational semantic analysis models, including: For any two service record data, multiple preset standard student record data are input into the educational semantic analysis model corresponding to the two service record data to obtain the analysis results of the two educational semantic analysis models for each standard student record data. For each standard student record data, calculate the similarity of the results of each evaluation dimension between the analysis results of the two educational semantic analysis models for that standard student record data. The rating dimensions with similarity scores higher than the preset similarity threshold are selected to obtain the set of associated dimensions corresponding to the student record data of this standard. Calculate the intersection of the sets of associated dimensions of all calibration student record data to obtain the set of evaluation dimensions corresponding to the two service record data; All evaluation dimensions in the evaluation dimension set corresponding to the two service record data are identified as association dimensions, and the semantic matching rules corresponding to the two service record data are obtained.

[0087] As can be seen, through the above optional embodiments, by inputting standard student record data into the educational semantic analysis model corresponding to two service records, calculating the similarity of the evaluation dimension results, filtering the set of related dimensions, and taking the intersection to obtain semantic matching rules, the automatic extraction of cross-model semantic association rules based on standard samples is realized, improving the objectivity and consistency of the rules, and reducing the risk of matching deviation caused by the subjectivity of manual rules.

[0088] As an optional embodiment, the fusion module performs fusion processing on all record analysis results according to semantic matching rules to obtain the multidimensional data analysis results corresponding to the student objects in the following specific ways: For any two record analysis results, cross-dimensional correction is performed on the two record analysis results based on semantic matching rules to obtain the corrected dimension results; Record all the results of the correction dimensions and the process of cross-dimensional correction, and identify them as the analysis data corresponding to the student subjects; The analysis data is input into the trained LLM model to obtain the performance evaluation of the student object in multiple dimensions; the LLM model is trained on a training dataset that includes multiple training analysis data and corresponding performance evaluation labels.

[0089] As can be seen, through the above optional embodiments, by performing cross-dimensional correction on the record analysis results based on semantic matching rules to obtain the corrected dimension results, and inputting them into the LLM model to output multi-dimensional performance evaluation, a precise comprehensive evaluation of student performance is achieved by integrating the correlation dimension correction with the large model, thereby improving the interpretability and comprehensiveness of the evaluation results and reducing the risk of biased performance evaluation caused by uncorrected dimensions.

[0090] As an optional embodiment, the fusion module performs cross-dimensional correction on the analysis results of the two records based on semantic matching rules to obtain the specific method of corrected dimensional results, including: For each evaluation dimension, based on the semantic matching rules between the service record data corresponding to the analysis results of the two records, it is determined whether the evaluation dimension is a related dimension, and the first judgment result is obtained; If the first judgment result is yes, calculate the weighted sum of the evaluation data of the two record analysis results in that evaluation dimension to obtain the correction dimension result corresponding to that evaluation dimension; If the first judgment result is negative, calculate the proportion of the number of pre-set similarity dimensions corresponding to the evaluation dimension that belong to the related dimensions in the corresponding semantic matching rules; Determine whether the quantity ratio is greater than a preset ratio threshold to obtain a second determination result; If the second judgment result is yes, calculate the weighted sum of the evaluation data of the two record analysis results in that evaluation dimension to obtain the correction dimension result corresponding to that evaluation dimension; If the second judgment result is negative, stop the correction step for that evaluation dimension.

[0091] As can be seen, through the above optional embodiments, by determining whether the evaluation dimension is a related dimension or the proportion of similar dimensions exceeds the threshold, a weighted summation correction is performed. This achieves accurate dimension correction based on both rules and proportions, improves the pertinence and robustness of the correction process, and reduces the risk of over- or under-correction caused by misjudgment of dimension association.

[0092] Example 3 Please see Figure 3 , Figure 3 This is another student multidimensional data management system based on cross-model semantic fusion disclosed in the embodiments of the present invention. Figure 3 The described student multidimensional data management system based on cross-model semantic fusion is applied in a data processing system / data processing equipment / data processing server (wherein, the server includes a local processing server or a cloud processing server). Figure 3 As shown, the student multidimensional data management system based on cross-model semantic fusion may include: Memory 301 storing executable program code; Processor 302 coupled to memory 301; The processor 302 calls the executable program code stored in the memory 301 to execute the steps of the student multidimensional data management method based on cross-model semantic fusion described in Embodiment 1.

[0093] Example 4 This invention discloses a computer read storage medium that stores a computer program for electronic data interchange, wherein the computer program causes a computer to execute the steps of the student multidimensional data management method based on cross-model semantic fusion described in Embodiment 1.

[0094] Example 5 This invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform the steps of the student multidimensional data management method based on cross-model semantic fusion described in Embodiment 1.

[0095] The foregoing has described specific embodiments of this specification; other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily have to follow the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0096] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0097] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware components.

[0098] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0099] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0100] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0101] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0102] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0103] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0104] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0105] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0106] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0107] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0108] Finally, it should be noted that the student multidimensional data management method and system based on cross-model semantic fusion disclosed in the embodiments of the present invention are merely preferred embodiments of the present invention and are only used to illustrate the technical solutions of the present invention, not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for managing multidimensional student data based on cross-model semantic fusion, characterized in that, The method includes: Acquire service record data from multiple dimensions generated by student objects across various educational service scenarios; Each service record data is input into the corresponding educational semantic analysis model of the educational service scenario to obtain the corresponding record analysis results; Based on the model matching between multiple educational semantic analysis models, determine the semantic matching rules between any two service record data; The analysis results of all the records are fused according to the semantic matching rules to obtain the multidimensional data analysis results corresponding to the student object.

2. The student multidimensional data management method based on cross-model semantic fusion according to claim 1, characterized in that, The service record data includes at least one of visual data, sound data, tactile data, light reflection data, text data, and computer operation data.

3. The student multidimensional data management method based on cross-model semantic fusion according to claim 1, characterized in that, The educational service scenarios include classroom service scenarios, self-study service scenarios, exercise service scenarios, examination service scenarios, or extracurricular activity service scenarios.

4. The student multidimensional data management method based on cross-model semantic fusion according to claim 3, characterized in that, The step of inputting each service record data into the corresponding educational semantic analysis model of the educational service scenario to obtain the corresponding record analysis results includes: For each service record data, determine the educational semantic analysis model corresponding to the educational service scenario for that service record data; the educational semantic analysis model includes a data vectorization model and a semantic analysis model. The service record data is input into the data vectorization model to obtain the output semantic relevance vector; The semantic relevance vectors are input into the semantic analysis model to obtain the record analysis results corresponding to the service record data; the record analysis results include quantitative data corresponding to multiple evaluation dimensions of the corresponding education service scenario; the semantic analysis model is trained using a training dataset that includes multiple training service record data in the corresponding education service scenario, as well as corresponding semantic relevance vector annotations and multi-dimensional analysis annotations.

5. The student multidimensional data management method based on cross-model semantic fusion according to claim 4, characterized in that, The data vectorization model includes a data semantic prediction network and a vectorized search network; the data semantic prediction network is used to predict the semantic content corresponding to the service record data; the vectorized search network is used to perform the following steps: Calculate the first similarity between the historical processing data and the service record data corresponding to each candidate vectorization model; the candidate vectorization model is trained using a training dataset that includes training record data from multiple educational service scenarios, as well as corresponding semantic content annotations and semantically related vector annotations; The semantic content and the preset template record data are input into each of the candidate vectorization models to obtain the corresponding semantic output vector; Calculate the second similarity between the historical output semantic vector and the semantic output vector corresponding to each candidate vectorized model; Calculate the product of the first similarity and the second similarity to obtain the model matching degree of each candidate vectorized model; The candidate vectorized model with the highest model matching degree is determined as the target vectorized model; The semantic content and the service record data are input into the target vectorization model to obtain the output semantically related vector.

6. The student multidimensional data management method based on cross-model semantic fusion according to claim 1, characterized in that, The step of determining the semantic matching rule between any two service record data based on model matching among multiple educational semantic analysis models includes: For any two service record data, multiple preset standard student record data are input into the educational semantic analysis model corresponding to the two service record data to obtain the analysis results of the two educational semantic analysis models for each standard student record data; For each of the standard student record data, calculate the similarity of the results for each evaluation dimension between the analysis results of the two educational semantic analysis models for the standard student record data; The rating dimensions with similarity scores higher than a preset similarity threshold are selected to obtain the set of associated dimensions corresponding to the standard student record data; Calculate the intersection of the associated dimension sets of all the calibration student record data to obtain the evaluation dimension set corresponding to the two service record data; All evaluation dimensions in the evaluation dimension set corresponding to the two service record data are determined as association dimensions, and the semantic matching rules corresponding to the two service record data are obtained.

7. The student multidimensional data management method based on cross-model semantic fusion according to claim 6, characterized in that, The step of fusing all the record analysis results according to the semantic matching rules to obtain the multidimensional data analysis results corresponding to the student object includes: For any two record analysis results, cross-dimensional correction is performed on the two record analysis results based on the semantic matching rules to obtain the corrected dimension result; Record all the results of the correction dimensions and the process of the cross-dimensional correction, and determine them as the analysis data corresponding to the student object; The analysis data is input into the trained LLM model to obtain the performance evaluation of the student object in multiple dimensions; the LLM model is trained using a training dataset that includes multiple training analysis data and corresponding performance evaluation labels.

8. The student multidimensional data management method based on cross-model semantic fusion according to claim 7, characterized in that, The cross-dimensional correction of the two record analysis results based on the semantic matching rules to obtain the corrected dimensional result includes: For each evaluation dimension, based on the semantic matching rules between the service record data corresponding to the two record analysis results, it is determined whether the evaluation dimension is a related dimension, and a first judgment result is obtained; If the first judgment result is yes, calculate the weighted sum of the evaluation data of the two record analysis results in the evaluation dimension to obtain the correction dimension result corresponding to the evaluation dimension; If the first judgment result is negative, calculate the proportion of the number of multiple preset similarity dimensions corresponding to the evaluation dimension that belong to the associated dimensions in the corresponding semantic matching rule; Determine whether the quantity ratio is greater than a preset ratio threshold to obtain a second determination result; If the second judgment result is yes, calculate the weighted sum of the evaluation data of the two recorded analysis results in the evaluation dimension to obtain the correction dimension result corresponding to the evaluation dimension; If the second judgment result is negative, stop the correction step for that evaluation dimension.

9. A student multidimensional data management system based on cross-model semantic fusion, characterized in that, The system includes: The acquisition module is used to acquire multi-dimensional service record data generated by student objects in multiple educational service scenarios; The analysis module is used to input the service record data of each service into the educational semantic analysis model of the corresponding educational service scenario to obtain the corresponding record analysis results. The determination module is used to determine the semantic matching rules between any two service record data based on the model matching between multiple educational semantic analysis models; The fusion module is used to fuse all the record analysis results according to the semantic matching rules to obtain the multidimensional data analysis results corresponding to the student object.

10. A student multidimensional data management system based on cross-model semantic fusion, characterized in that, The system includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the student multidimensional data management method based on cross-model semantic fusion as described in any one of claims 1-8.

Images