A dynamic evaluation method and system of a classroom interaction network

Abstract

This application relates to the field of smart education technology and discloses a dynamic evaluation method and system integrating a classroom interaction network, comprising: an interaction data acquisition module for collecting interaction feature data in the classroom; a classroom interaction network construction module, which adopts a multimodal feature mapping and dynamic network update fusion architecture to perform node attribute association, edge type matching, and weight quantification calculation on the interaction feature data to construct a three-dimensional network model of nodes, edges, and weights; a dynamic evaluation module, which constructs a three-level dynamic evaluation model based on the dynamic features of the three-dimensional network model to dynamically evaluate classroom teaching, teacher-student interaction, and student learning status; and a data storage module for storing full-process data and enabling data sharing with the teaching management system; effectively improving the real-time and comprehensiveness of classroom evaluation, while providing data support for optimizing teaching strategies and personalized learning guidance.

Description

Technical Field

[0001] This specification relates to the field of smart education technology, and in particular to a dynamic evaluation method and system that integrates classroom interactive networks. Background Technology

[0002] With the advancement of digital transformation in education, classroom interaction has become increasingly diverse, evolving from traditional one-way question-and-answer sessions between teachers and students to multi-way interactions between teachers and students, and among students themselves, encompassing various forms such as voice dialogue, instant quizzes, group discussions, and interactive comments. The quality of classroom interaction directly impacts teaching effectiveness and student learning outcomes; therefore, scientific and accurate evaluation of classroom interaction is crucial for improving teaching quality.

[0003] Existing classroom assessment methods suffer from the following shortcomings: First, the assessment methods are static, relying heavily on post-lesson debriefing, student questionnaire feedback, or single-dimensional data analysis (such as answer accuracy). They fail to capture the dynamic changes during classroom interaction in real time, resulting in significant assessment lag and making it difficult to meet the needs of real-time adjustments in classroom teaching. Second, the assessment dimensions are limited. Existing technologies often focus on evaluating a single form of interaction (such as voice interaction analysis based solely on acoustic signals), failing to integrate multiple interaction forms into a complete classroom interaction network. This results in an incomplete reflection of the overall structure, correlation strength, and individual participation differences in classroom interaction, leading to biased assessment results that are difficult to support precise teaching optimization. Third, the interaction network is disconnected from the assessment system. The few existing technologies involving classroom interaction networks can only achieve statistical analysis and network construction of interactive behaviors, failing to deeply integrate network characteristics with teaching evaluation indicators. They cannot quantify interaction quality through network characteristics and lack dynamic assessment capabilities and targeted optimization suggestions.

[0004] Therefore, there is an urgent need for a dynamic evaluation method and system that integrates classroom interactive networks to overcome the aforementioned problems of existing technologies. Summary of the Invention

[0005] In view of this, the present invention aims to propose a dynamic evaluation method and system that integrates classroom interactive networks to solve the problems of lagging evaluation, single dimension, and one-sided results caused by the lack of multimodal interactive data fusion mechanism, dynamic network modeling capability, and deep coupling design of network characteristics and evaluation system in existing classroom evaluation systems. This invention aims to achieve accurate dynamic evaluation and efficient data sharing of classroom teaching, teacher-student interaction and student learning status.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A dynamic evaluation system integrating classroom interactive networks, comprising: The interactive data acquisition module is used to collect voice interaction data, text interaction data, and behavioral interaction data in the classroom; and to clean, standardize, and extract features from the collected data through a built-in standardized interface to obtain standardized interactive feature data. The classroom interaction network construction module adopts a fusion architecture of multimodal feature mapping and dynamic network update. It performs node attribute association, edge type matching and weight quantization calculation on the interaction feature data output by the interaction data acquisition module to construct a three-dimensional network model of nodes, edges and weights. The dynamic evaluation module, based on the real-time dynamic features of the three-dimensional network model of the classroom interaction network construction module, constructs a three-level dynamic evaluation model to dynamically evaluate classroom teaching, teacher-student interaction, and student learning status; the real-time dynamic features of the three-dimensional network model include node degree, network density, interaction weight, and emotion conversion rate. The data storage module uses a distributed database to store all process data and enables data sharing with the teaching management system based on a standardized data interface.

[0007] The beneficial effects of this solution are as follows: In existing technologies, classroom evaluation systems generally suffer from static and outdated evaluation methods, limited interactive data dimensions, and a disconnect between evaluation and network characteristics, which restricts the accuracy of classroom evaluation and its value in guiding teaching. This system, through multimodal data fusion and standardized processing in the interactive data acquisition module, combined with the three-dimensional network modeling capabilities of the classroom interactive network construction module, and through the deep coupling of the three-level evaluation model of the dynamic evaluation module with network characteristics, achieves real-time, comprehensive, and accurate dynamic evaluation of classroom teaching, teacher-student interaction, and student learning status. Furthermore, through end-to-end data sharing in the data storage module, it provides reliable support for teaching optimization, effectively improving the timeliness and teaching application value of classroom evaluation.

[0008] Furthermore, the interactive feature data includes fundamental frequency features, amplitude energy construction features, emotion features, content relevance features, participation features, interaction weight features, and classroom emotion conversion rate features.

[0009] Beneficial effects: By extracting multi-dimensional interaction features, richer and more accurate basic data support is provided for the construction of subsequent classroom interaction networks and dynamic evaluation, effectively improving the comprehensiveness and accuracy of evaluation results.

[0010] Furthermore, the standardized interface built into the interactive data acquisition module includes data modality identifier and data verification code as interface parameters; the data modality identifier adopts 3-bit binary encoding, and the data verification code is generated by combining the acquired data subject and the timestamp using the SHA-256 algorithm.

[0011] Beneficial effects: Through the design of standardized interface data modality identification and SHA-256 checksum, it not only achieves accurate classification and efficient transmission of multimodal interactive data, but also ensures the integrity and immutability of the collected data, providing a reliable foundation for subsequent data processing and network modeling.

[0012] Furthermore, in the three-dimensional network model of the classroom interaction network construction module, the weights are calculated based on interaction frequency, interaction depth, and interaction effectiveness, and the calculation logic is as follows: ; In the formula, Interaction frequency For interaction depth, For the validity of the interaction; , , The weighting coefficients and .

[0013] Beneficial Effects: By designing the weights as a weighted combination of interaction frequency, interaction depth, and interaction effectiveness, this approach overcomes the limitations of measuring interaction quality solely based on frequency, enabling a more precise quantification of the actual value of different interactive behaviors. This calculation logic considers both the activity level and the depth and effectiveness of the interaction, providing a more practical quantitative basis for the subsequent construction of the three-dimensional network model and dynamic evaluation, effectively improving the accuracy and rationality of classroom interaction evaluation.

[0014] Furthermore, the classroom interaction network construction module determines the interaction depth. The process includes: When the interaction data is voice interaction data, a dialogue duration of 30 seconds or more is considered a deep interaction, and less than 30 seconds is considered a shallow interaction; when the interaction data is text interaction data, targeted comments related to the teaching content are considered a deep interaction, and meaningless comments are considered a shallow interaction. When the interaction data is behavioral interaction data, it is judged as deep interaction if the group participation time is greater than or equal to 5 minutes, the number of screen sharing times is greater than or equal to 2 times, or the number of key content marks is greater than or equal to 3 times. It is judged as shallow interaction if the person raises their hand to respond but does not actually participate or the group participation time is less than 2 minutes.

[0015] Beneficial effects: By establishing clear depth judgment rules for three types of interactive data—voice, text, and behavior—objective quantification of interaction depth is achieved, avoiding the bias of subjective judgment. This makes the weight calculation of the 3D network model more closely match the actual interaction quality in classroom teaching scenarios, providing accurate and traceable deep interaction data support for subsequent dynamic evaluation.

[0016] Furthermore, the classroom interaction network construction module determines the effectiveness of the interaction. Based on semantic similarity and teacher feedback results, a lightweight attention fusion model was used to obtain the results; the steps are as follows: Extract semantic feature vectors of interactive content Obtain teacher comment tag vector ; Attention weights for semantic features are calculated using a single-layer perceptron. The calculation logic is as follows: ; In the formula, It is a 128×1 dimensional weight matrix. For bias terms, Use the Sigmoid activation function; The semantic features and comment tags are fused using a weighted summation method to obtain the fused features. The calculation logic is as follows: ; In the formula, This is the normalized result of the label vector; It is a semantic feature vector; Fusion features Input a logistic regression classifier and output a quantified value of the interaction effectiveness. The calculation logic is as follows: ; In the formula, For the corresponding dimension weight matrix, For bias terms, ∈[0,1], A value ≥0.6 is considered a valid interaction. A value less than 0.6 is considered an invalid interaction.

[0017] Beneficial effects: By integrating semantic similarity with teacher feedback results and combining a lightweight attention fusion model to quantitatively calculate the effectiveness of interactions, this approach relies on semantic features to objectively determine the relevance of interactive content to teaching while incorporating professional teaching judgments from teacher feedback, making the determination of interaction effectiveness more aligned with actual teaching practices. Simultaneously, the lightweight model design balances computational efficiency and accuracy, enabling rapid output of quantitative results. This provides an objective and accurate basis for calculating the weights of the 3D network model, further enhancing the scientific rigor of classroom interaction evaluation.

[0018] Furthermore, the process of constructing the three-level dynamic evaluation model by the dynamic evaluation module includes: The process evaluation is performed, the dynamic features of the interactive network are extracted in real time, the real-time values ​​of various evaluation indicators are calculated, the process evaluation results are output every 5 minutes and the interaction anomalies are identified; the interaction anomalies include network density that is too low and the degree of individual student nodes being 0. The phased evaluation is implemented, dividing each lesson into three phases: pre-class preparation, in-class teaching, and post-class summary. After each phase, based on the interactive network characteristics of that phase and combined with the phase teaching objectives, the phase evaluation score is calculated, the matching degree between the interactive effect and the teaching objectives of that phase is analyzed, and the problems existing in that phase are identified. The system performs real-time feedback evaluation. When an evaluation indicator shows an anomaly, real-time feedback is automatically triggered, and the cause of the anomaly is analyzed in conjunction with the characteristics of the interactive network.

[0019] Beneficial effects: By constructing a three-level dynamic evaluation model that combines process-oriented, phased, and real-time feedback, the model achieves real-time, multi-dimensional dynamic evaluation of classroom interaction. It can capture dynamic changes in classroom interaction in real time and identify anomalies, conduct targeted evaluations in conjunction with phased teaching objectives, and automatically trigger feedback and analyze the causes of abnormal evaluation indicators. This effectively solves the problems of lagging traditional classroom evaluation and lack of real-time guidance, and provides a precise basis for real-time adjustment and phased optimization of classroom teaching.

[0020] Furthermore, the dynamic evaluation module's execution of the dynamic evaluation process also includes evaluation score calculation, including: The evaluation metrics are standardized, including node degree, network density, and interaction weight. The real-time value of the emotion conversion rate is mapped to the [0,1] interval using a min-max algorithm; the calculation logic is as follows: ; In the formula, The real-time value of the evaluation index; The minimum value of the evaluation index; The maximum value of the evaluation index; Based on interaction weights The calculation logic determines the weight of each evaluation indicator. The sum of the weights is 1. The evaluation score is calculated using a weighted summation algorithm. Its calculation logic is as follows: ; In the formula, To evaluate the score; Weights for each evaluation indicator; These are the standardized values ​​of each indicator.

[0021] Beneficial effects: By standardizing the evaluation indicators through the min-max algorithm, allocating indicator weights by combining interactive weight calculation logic, and quantifying the evaluation scores using a weighted summation algorithm, unified quantification and comprehensive scoring of evaluation indicators across different dimensions are achieved. This allows classroom evaluation results to be presented in an intuitive numerical form, ensuring the scientific and objective nature of the scoring calculation while improving the readability and comparability of the evaluation results. It provides accurate and unified numerical basis for the quantitative analysis and grade determination of classroom teaching evaluation.

[0022] Furthermore, a dynamic assessment method integrating classroom interactive networks includes: S1. Collect classroom voice, text and behavioral interaction data, clean, standardize and extract features from the collected data to obtain standardized interaction feature data; S2. Perform node attribute association, edge type matching and weight quantification calculation on standardized interactive feature data to construct a three-dimensional classroom interaction network model of nodes, edges and weights. S3. Based on the real-time dynamic characteristics of the three-dimensional classroom interactive network model, a three-level dynamic evaluation model is constructed and the evaluation is executed. The evaluation score is calculated synchronously, and finally the entire process data is stored and shared with the teaching management system.

[0023] Beneficial effects: This method is compatible with dynamic evaluation systems that integrate classroom interactive networks. It achieves a closed-loop process of interactive data acquisition and processing, 3D interactive network modeling, three-level dynamic evaluation, and data sharing through step-by-step steps. It translates the technical advantages of each module of the system into an executable method and process, which not only ensures the real-time, comprehensive, and accurate nature of classroom evaluation, but also achieves efficient storage and sharing of evaluation data. The process logic is clear and the operability is strong, which can effectively support the dynamic evaluation and teaching optimization of smart classrooms. Attached Figure Description

[0024] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein: Figure 1 This is an exemplary structural diagram of a dynamic evaluation system that integrates classroom interactive networks; Figure 2 This is an exemplary flowchart of a dynamic assessment method that integrates classroom interactive networks. Detailed Implementation

[0025] To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0026] As indicated in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0027] Flowcharts are used in this specification to illustrate the operations performed by the system according to embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0028] The following detailed explanation illustrates the specific implementation methods: Example 1: Figure 1 This is an exemplary structural diagram of a dynamic evaluation system that integrates classroom interactive networks, such as... Figure 1 As shown, a dynamic evaluation system integrating a classroom interaction network (hereinafter referred to as the "classroom interaction system") includes: The interactive data acquisition module is used to collect voice interaction data, text interaction data, and behavioral interaction data in the classroom; and to clean, standardize, and extract features from the collected data through a built-in standardized interface to obtain standardized interactive feature data.

[0029] Voice interaction data refers to the data related to interactions generated between teachers and students through voice dialogue during classroom teaching. In this embodiment, voice interaction data includes speaker identity, voice duration, emotional state, and interaction trigger time.

[0030] Text interaction data refers to the data generated during classroom teaching through text-based communication and interaction between teachers and students. In this embodiment, text interaction data includes answers, comments, questions, and related interaction information.

[0031] Behavioral interaction data refers to the interaction-related data generated by the actions and behaviors of teachers and students during classroom teaching. In this embodiment, behavioral interaction data includes the number of times hands are raised, the duration of group participation, the number of times screen sharing is performed, and interactive guidance behaviors.

[0032] In this embodiment, the classroom interaction system can collect various types of interaction data in multiple ways.

[0033] For example, classroom interaction systems can collect voice interaction data through recording devices; text interaction data between teachers and students can be collected through online teaching platforms; and student behavior interaction data can be collected through smart terminals or classroom monitoring devices. This enables the comprehensive collection of classroom interaction data across multiple scenarios and types, providing a complete data foundation for subsequent data processing and network modeling.

[0034] Interaction feature data is core data that can characterize the essential attributes of classroom interaction. In this embodiment, interaction feature data includes fundamental frequency features, amplitude energy construction features, emotion features, content relevance features, participation features, interaction weight features, and classroom emotion conversion rate features.

[0035] In this embodiment, the classroom interaction system cleans, standardizes, and extracts features from the collected data through a built-in standardized interface to obtain the interaction feature data.

[0036] The built-in standardized interface includes data modality identifiers and data verification codes as interface parameters.

[0037] The data modality identifier uses a 3-bit binary code, and the data check code is generated by combining the collected data subject and the timestamp using the SHA-256 algorithm.

[0038] As an example, a 3-bit binary code is used to identify voice interaction data, text interaction data, and behavioral interaction data: "001" for voice interaction data, "010" for text interaction data, and "100" for behavioral interaction data. This enables rapid and accurate classification and differentiation of multimodal data. Simultaneously, the SHA-256 algorithm is used to combine the original interaction data with the corresponding timestamp to generate a unique data verification code. This effectively verifies whether the data has been tampered with or lost during transmission and processing, ensuring the integrity and reliability of the collected data and providing a guarantee for the accurate extraction of subsequent interaction feature data.

[0039] The classroom interaction network construction module adopts a fusion architecture of multimodal feature mapping and dynamic network update. It performs node attribute association, edge type matching and weight quantization calculation on the interaction feature data output by the interaction data acquisition module to construct a three-dimensional network model of nodes, edges and weights.

[0040] Among them, multimodal feature mapping refers to converting the interactive feature data of three different modalities—voice, text, and behavior—output by the interactive data acquisition module into a unified dimensional feature vector that can be used for network modeling through a unified feature mapping rule. This achieves the normalization and correlation fusion of interactive features of different modalities, ensuring that multi-type interactive data can collaboratively support network construction. Dynamic network update refers to synchronously capturing newly generated interactive feature data as classroom interaction progresses in real time, and updating the node attributes, edge types, and weight values ​​in the three-dimensional network model in real time. This enables the constructed classroom interactive network to accurately reflect the dynamic changes in classroom interaction and avoids the network model from becoming disconnected from the actual classroom interaction scenario.

[0041] In the three-dimensional network model of nodes, edges, and weights, nodes represent teachers or students, edges represent interactive behaviors, and weights represent interaction frequency, interaction depth, and interaction effectiveness.

[0042] Interaction frequency refers to the actual number of times various interactive behaviors, such as voice, text, and behavior, occur between teachers and students during classroom teaching.

[0043] Interaction depth refers to the degree of participation and content relevance achieved by classroom interaction behavior. It is a quantitative indicator that distinguishes between shallow and deep classroom interaction.

[0044] Interaction effectiveness refers to the degree to which the content of classroom interaction behaviors aligns with teaching objectives and their actual teaching value. It is a quantitative evaluation indicator of whether interactive behaviors have positive teaching significance.

[0045] Furthermore, in the three-dimensional network model of the classroom interaction network construction module, the weights are calculated based on interaction frequency, interaction depth, and interaction effectiveness. The calculation logic is as follows: ; In the formula, Interaction frequency For interaction depth, For the validity of the interaction; , , The weighting coefficients and .

[0046] In this embodiment, the frequency of interaction The actual number of voice, text, and physical interactions between teachers and students during classroom teaching can be directly counted through the classroom interaction system.

[0047] In this embodiment, the interaction depth The classroom interaction system achieves automated quantification by matching preset quantification rules with data thresholds. ∈[0,1].

[0048] Furthermore, the classroom interactive network construction module determines the depth of interaction. The process includes: When the interaction data is voice interaction data, a dialogue duration of 30 seconds or more is considered a deep interaction, and less than 30 seconds is considered a shallow interaction; when the interaction data is text interaction data, targeted comments related to the teaching content are considered a deep interaction, and meaningless comments are considered a shallow interaction.

[0049] When the interaction data is behavioral interaction data, it is judged as deep interaction if the group participation time is greater than or equal to 5 minutes, the number of screen sharing times is greater than or equal to 2 times, or the number of key content marks is greater than or equal to 3 times. It is judged as shallow interaction if the person raises their hand to respond but does not actually participate or the group participation time is less than 2 minutes.

[0050] The system backend pre-configures and embeds the interaction depth quantization threshold and corresponding numerical mapping rules into the program. When collected interaction data is input, the system automatically matches the preset threshold using a data comparison algorithm and outputs the corresponding interaction depth quantization value. Specifically, if the voice interaction data meets the above-mentioned deep interaction conditions, a value is assigned... =1; if the above shallow interaction conditions are met, assign a value. =0.3; If the above deep interaction conditions are met in the text interaction data, assign a value of 0.3. =1, if the above shallow interaction conditions are met, assign a value. =0.2; If the behavioral interaction data meets the above deep interaction conditions, assign a value of 0.2. =1; if the above shallow interaction conditions are met, assign a value. =0.4. When the same interactive behavior meets multiple deep interaction judgment conditions, the system automatically takes the highest quantization value and does not repeatedly add them together.

[0051] Furthermore, the classroom interaction network construction module determines the effectiveness of the interaction. Based on semantic similarity and teacher feedback results, a lightweight attention fusion model was used to obtain the results.

[0052] Semantic similarity refers to the degree of alignment between the content of classroom interactive behaviors (such as voice and text interaction) and the current classroom teaching objectives and content.

[0053] Teacher feedback results are quantitative values ​​of the evaluation opinions given by teachers for various interactive behaviors in the classroom.

[0054] Determine the validity of the interaction The steps are as follows: Step 11: Extract the semantic feature vector of the interactive content. Obtain teacher comment tag vector .

[0055] In this embodiment, semantic feature vector The extraction is achieved using the BERT pre-trained language model, which is a mature and publicly available model in the field of natural language processing. It can be directly adapted to the interactive content feature extraction scenario of this classroom interactive system without additional retraining.

[0056] Specifically, the classroom interaction content undergoes standardized preprocessing to remove meaningless interjections, punctuation marks, and redundant information, resulting in clean interactive text. This preprocessed text is then input into a pre-trained BERT model, where its feature extraction layer automatically captures semantic information and outputs a semantic feature vector with uniform dimensions. .

[0057] Teacher feedback label vector The acquisition is achieved by combining standardized mapping with existing tag encoding technology.

[0058] Specifically, the classroom interaction system pre-establishes mapping rules between teacher comments and quantitative labels. For each classroom interaction, the teacher selects the corresponding comment (e.g., excellent, good, satisfactory, invalid) through the comment entry point of the classroom interaction system. After receiving the teacher's comment, the system automatically converts the comment into a standardized quantitative value in the [0,1] range according to the preset mapping rules. Then, using existing label encoding technology, the quantitative value is converted into a dimensional and semantic feature vector. Consistent teacher feedback label vector This ensures that the two can be fused and calculated subsequently.

[0059] Step 12: Calculate the attention weights of semantic features using a single-layer perceptron. The calculation logic is as follows: ; In the formula, It is a 128×1 dimensional weight matrix. For bias terms, This is the Sigmoid activation function.

[0060] Step 13: Fuse semantic feature vectors using a weighted summation method. Teacher feedback label vector , obtain fusion features The calculation logic is as follows: ; In the formula, This is the normalized result of the label vector; This is a semantic feature vector.

[0061] Step 14, fuse features Input a logistic regression classifier and output a quantified value of the interaction effectiveness. The calculation logic is as follows: ; In the formula, For the corresponding dimension weight matrix, For bias terms, ∈[0,1], A value ≥0.6 is considered a valid interaction. A value less than 0.6 is considered an invalid interaction.

[0062] In this embodiment, the effectiveness of interaction is quantitatively calculated by integrating semantic similarity and teacher feedback results with a lightweight attention fusion model. This approach relies on semantic features to objectively determine the relevance of the interaction content to teaching, while also incorporating the professional teaching judgment of teacher feedback, making the determination of interaction effectiveness more aligned with actual teaching practices. At the same time, the lightweight model design balances computational efficiency and judgment accuracy, enabling rapid output of quantitative results. This provides an objective and accurate basis for the calculation of weights in the three-dimensional network model, further enhancing the scientific nature of classroom interaction evaluation.

[0063] The dynamic evaluation module, based on the real-time dynamic features of the three-dimensional network model of the classroom interaction network construction module, constructs a three-level dynamic evaluation model to dynamically evaluate classroom teaching, teacher-student interaction, and student learning status.

[0064] The real-time dynamic features of the 3D network model refer to the core feature set that changes in real time as the classroom teaching progresses, accurately reflecting the operational status of the 3D network model, namely nodes, edges, and weights, and can be used for dynamic classroom evaluation. In this embodiment, it is directly derived from the 3D network model output by the classroom interactive network construction module.

[0065] In this embodiment, the real-time dynamic features of the three-dimensional network model include node degree, network density, interaction weight, and sentiment conversion rate.

[0066] Among them, the node degree corresponds to the frequency of classroom interaction of each node (representing teachers and students) in the three-dimensional network model.

[0067] Network density corresponds to the density of interactive edges in a three-dimensional network model, reflecting the overall level of classroom interaction activity.

[0068] Interaction weights correspond to the quantified weights of each interaction edge in the 3D network model, reflecting the actual teaching value of different interaction behaviors.

[0069] The emotion conversion rate corresponds to the emotion features extracted from interaction data in the three-dimensional network model, reflecting the efficiency of positive transformation of teachers' and students' emotions during classroom interaction.

[0070] In this embodiment, the aforementioned features work together to form the core data support for dynamic classroom evaluation, ensuring that the three-level dynamic evaluation model can comprehensively, in real time, and accurately evaluate classroom teaching, teacher-student interaction, and student learning status.

[0071] Furthermore, the process of constructing a three-level dynamic evaluation model in the dynamic evaluation module includes: The process evaluation is performed, the dynamic features of the interactive network are extracted in real time, the real-time values ​​of various evaluation indicators are calculated, the process evaluation results are output every 5 minutes and the interaction anomalies are identified; the interaction anomalies include network density that is too low and the degree of individual student nodes being 0.

[0072] In this embodiment, the classroom interaction system captures the real-time changes of nodes, edges, and weights in the three-dimensional network model at the millisecond level through a real-time data synchronization interface, and automatically extracts four core dynamic features: node degree, network density, interaction weight, and emotion conversion rate.

[0073] Among them, the node degree is directly calculated by the number of associated edges of each teacher and student node in the statistical model; the network density is calculated by the ratio of the actual number of interaction edges in the model to the theoretical maximum number of interaction edges; the interaction weight directly retrieves the real-time quantitative calculation value of each edge in the model and takes the overall average as the indicator value; and the emotion conversion rate is calculated by combining the teacher and student emotion feature data extracted in the model with the correlation algorithm between the positive emotion ratio and the interaction behavior.

[0074] The implementation of phased evaluation involves dividing each lesson into three phases: pre-class preparation, in-class teaching, and post-class summary. After each phase, based on the interactive network characteristics of that phase and in conjunction with the phase's teaching objectives, a phased evaluation score is calculated. The matching degree between the interactive effect and the teaching objectives of that phase is analyzed, and problems existing in that phase are identified.

[0075] The system performs real-time feedback evaluation. When an evaluation indicator shows an anomaly, real-time feedback is automatically triggered, and the cause of the anomaly is analyzed in conjunction with the characteristics of the interactive network.

[0076] Furthermore, the dynamic evaluation module's execution of the dynamic evaluation process also includes evaluation score calculation, including: Step 21: Standardize the evaluation metrics by mapping the real-time values ​​of node degree, network density, interaction weight, and sentiment conversion rate to the [0,1] interval using a min-max algorithm; the calculation logic is as follows: ; In the formula, This refers to the real-time value of a specific evaluation indicator, such as the current network density in the classroom. This is the minimum value of the indicator within historical data or a preset range; This is the maximum value of the indicator within historical data or a preset range. It is obtained through calculation. This standardized value eliminates the numerical differences between different indicators, making subsequent calculations easier.

[0077] Step 22: Based on the calculation logic of interactive weights, determine the weights of each evaluation index, ensuring that the sum of the weights is 1.

[0078] Step 23: Calculate the evaluation score using a weighted summation algorithm. Its calculation logic is as follows: ; In the formula, To evaluate the score; Weights for each evaluation indicator; These are the standardized values ​​of each indicator.

[0079] In this embodiment, the evaluation indicators are standardized through the min-max algorithm. The indicator weights are allocated by the interactive weight calculation logic, and the evaluation scores are quantified by the weighted summation algorithm. This achieves unified quantification and comprehensive scoring of evaluation indicators of different dimensions, allowing classroom evaluation results to be presented in an intuitive numerical form. This ensures the scientificity and objectivity of the scoring calculation, while also improving the readability and comparability of the evaluation results. It provides accurate and unified numerical basis for the quantitative analysis and grade determination of classroom teaching evaluation.

[0080] The data storage module uses a distributed database to store all process data and enables data sharing with the teaching management system based on a standardized data interface.

[0081] In this embodiment, the module is used to store the raw data collected by the interactive data acquisition module, the preprocessed feature data, the dynamic data of the classroom interactive network, the dynamic evaluation results and historical data, and adopts distributed database storage to support fast data query, update and backup, ensuring data security and integrity. At the same time, it supports data docking with the existing teaching management system to achieve data sharing.

[0082] Example 2: Based on Example 1, further, as follows: Figure 2 As shown, a dynamic evaluation method integrating a classroom interactive network, applied to the aforementioned dynamic evaluation system integrating a classroom interactive network, includes: Step S1: Collect classroom voice, text, and behavioral interaction data; clean, standardize, and extract features from the collected data to obtain standardized interaction feature data.

[0083] Step S2 involves performing node attribute association, edge type matching, and weight quantization calculations on the standardized interactive feature data to construct a three-dimensional classroom interaction network model consisting of nodes, edges, and weights.

[0084] Step S3: Based on the real-time dynamic characteristics of the three-dimensional classroom interactive network model, construct a three-level dynamic evaluation model and perform the evaluation, simultaneously calculate the evaluation score, and finally store the entire process data and achieve data sharing with the teaching management system.

[0085] In this embodiment, the method is highly compatible with the corresponding dynamic evaluation system. By implementing the entire process of classroom interaction data collection and processing, 3D interactive network modeling, three-level dynamic evaluation, and data sharing in steps, it effectively improves the real-time, comprehensiveness, and accuracy of classroom evaluation, solving the problems of single-dimensionality, strong subjectivity, and low data utilization in traditional classroom evaluation. Its standardized data processing ensures the reliability of the evaluation basis, 3D network modeling accurately captures the core features of teacher-student interaction, and three-level evaluation and quantitative score calculation realize objective evaluation of classroom teaching, teacher-student interaction, and student learning status. At the same time, through data sharing and linkage with the teaching management system, it provides reliable methodological support for optimizing teaching strategies. The process is simple and operable, easy to implement and apply, and further contributes to improving the quality of smart classroom teaching and management efficiency.

[0086] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this specification. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification and therefore remain within the spirit and scope of the exemplary embodiments described herein.

[0087] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this specification are not intended to limit the order of the processes and methods described herein. Although various examples have been discussed in the foregoing disclosure of some embodiments of the invention that are currently considered useful, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the spirit and scope of the embodiments described herein. For example, while the system components described above can be implemented using hardware devices, they can also be implemented solely using software solutions, such as installing the described system on existing servers or mobile devices.

[0088] Similarly, it should be noted that, in order to simplify the description disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments in this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.

[0089] In some embodiments, numbers describing the quantity of components and attributes are used. It should be understood that such numbers used in the description of embodiments are modified in some examples with the terms "approximately," "approximately," or "generally." Unless otherwise stated, "approximately," "approximately," or "generally" indicates that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may be changed depending on the characteristics required by individual embodiments. In some embodiments, numerical parameters should take into account specified significant digits and employ a general method of digit reservation. Although the numerical ranges and parameters used to confirm their breadth of range in some embodiments of this specification are approximate values, in specific embodiments, such values ​​are set as precisely as feasible.

[0090] For each patent, patent application, patent application publication, and other material such as articles, books, specifications, publications, and documents referenced in this specification, the entire contents of which are incorporated herein by reference. This excludes historical application documents that are inconsistent with or conflict with the content of this specification, as well as documents that limit the broadest scope of the claims in this specification (currently or subsequently appended to this specification). It should be noted that in the event of any inconsistency or conflict between the descriptions, definitions, and / or terminology used in the supplementary materials to this specification and the content of this specification, the descriptions, definitions, and / or terminology used in this specification shall prevail.

[0091] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A dynamic evaluation system integrating classroom interactive networks, characterized in that, include: The interactive data acquisition module is used to collect voice interaction data, text interaction data, and behavioral interaction data in the classroom; and to clean, standardize, and extract features from the collected data through a built-in standardized interface to obtain standardized interactive feature data. The classroom interaction network construction module adopts a fusion architecture of multimodal feature mapping and dynamic network update. It performs node attribute association, edge type matching and weight quantization calculation on the interaction feature data output by the interaction data acquisition module to construct a three-dimensional network model of nodes, edges and weights. The dynamic evaluation module, based on the real-time dynamic features of the three-dimensional network model of the classroom interaction network construction module, constructs a three-level dynamic evaluation model to dynamically evaluate classroom teaching, teacher-student interaction, and student learning status. The real-time dynamic features of the three-dimensional network model include node degree, network density, interaction weight, and sentiment conversion rate. The data storage module uses a distributed database to store all process data and enables data sharing with the teaching management system based on a standardized data interface.

2. The dynamic evaluation system integrating classroom interactive networks according to claim 1, characterized in that, The interactive feature data includes fundamental frequency features, amplitude energy construction features, emotion features, content relevance features, participation features, interaction weight features, and classroom emotion conversion rate features.

3. The dynamic evaluation system integrating classroom interactive networks according to claim 1, characterized in that, The interactive data acquisition module has a built-in standardized interface with parameters including a data modality identifier and a data verification code. The data modality identifier uses a 3-bit binary code, and the data verification code is generated by combining the acquired data subject and the timestamp using the SHA-256 algorithm.

4. The dynamic evaluation system integrating classroom interactive networks according to claim 3, characterized in that, In the three-dimensional network model of the classroom interaction network construction module, the weights are calculated based on interaction frequency, interaction depth, and interaction effectiveness. The calculation logic is as follows: ； In the formula, Interaction frequency For interaction depth, For the validity of the interaction; , , The weighting coefficients and .

5. A dynamic evaluation system integrating a classroom interactive network according to claim 4, characterized in that, The classroom interaction network construction module determines the depth of interaction. The process includes: When the interaction data is voice interaction data, a dialogue duration of 30 seconds or more is considered a deep interaction, and less than 30 seconds is considered a shallow interaction; when the interaction data is text interaction data, targeted comments related to the teaching content are considered a deep interaction, and meaningless comments are considered a shallow interaction. When the interaction data is behavioral interaction data, it is judged as deep interaction if the group participation time is greater than or equal to 5 minutes, the number of screen sharing times is greater than or equal to 2 times, or the number of key content marks is greater than or equal to 3 times. It is judged as shallow interaction if the person raises their hand to respond but does not actually participate or the group participation time is less than 2 minutes.

6. The dynamic evaluation system integrating classroom interactive networks according to claim 4, characterized in that, The classroom interaction network construction module determines the effectiveness of the interaction. Based on semantic similarity and teacher feedback results, a lightweight attention fusion model was used to obtain the results; the steps are as follows: Extract semantic feature vectors of interactive content Obtain teacher comment tag vector ; Attention weights for semantic features are calculated using a single-layer perceptron. The calculation logic is as follows: ； In the formula, It is a 128×1 dimensional weight matrix. For bias terms, Use the Sigmoid activation function; The semantic features and comment tags are fused using a weighted summation method to obtain the fused features. The calculation logic is as follows: ； In the formula, This is the normalized result of the label vector; It is a semantic feature vector; Fusion features Input a logistic regression classifier and output a quantified value of the interaction effectiveness. The calculation logic is as follows: ； In the formula, For the corresponding dimension weight matrix, For bias terms, ∈[0,1], A value ≥0.6 is considered a valid interaction. A value less than 0.6 is considered an invalid interaction.

7. A dynamic evaluation system integrating a classroom interactive network according to claim 6, characterized in that, The process of constructing the three-level dynamic evaluation model by the dynamic evaluation module includes: The process evaluation is performed, the dynamic features of the interactive network are extracted in real time, the real-time values ​​of various evaluation indicators are calculated, the process evaluation results are output every 5 minutes and the interaction anomalies are identified; the interaction anomalies include network density that is too low and the degree of individual student nodes being 0. The phased evaluation is implemented, dividing each lesson into three phases: pre-class preparation, in-class teaching, and post-class summary. After each phase, based on the interactive network characteristics of that phase and combined with the phase teaching objectives, the phase evaluation score is calculated, the matching degree between the interactive effect and the teaching objectives of that phase is analyzed, and the problems existing in that phase are identified. The system performs real-time feedback evaluation. When an evaluation indicator shows an anomaly, real-time feedback is automatically triggered, and the cause of the anomaly is analyzed in conjunction with the characteristics of the interactive network.

8. A dynamic evaluation system integrating a classroom interactive network according to claim 7, characterized in that, The dynamic evaluation module also includes calculating evaluation scores during the dynamic evaluation process, including: The evaluation metrics are standardized, including node degree, network density, and interaction weight. The real-time value of the emotion conversion rate is mapped to the [0,1] interval using a min-max algorithm; the calculation logic is as follows: ； In the formula, The real-time value of the evaluation index; The minimum value of the evaluation index; The maximum value of the evaluation index; Based on interaction weights The calculation logic determines the weight of each evaluation indicator. The sum of the weights is 1. The evaluation score is calculated using a weighted summation algorithm. Its calculation logic is as follows: ； In the formula, To evaluate the score; Weights for each evaluation indicator; These are the standardized values ​​of each indicator.

9. A dynamic evaluation method integrating a classroom interactive network, applied to the dynamic evaluation system integrating a classroom interactive network as described in claim 1, characterized in that, include: S1. Collect classroom voice, text and behavioral interaction data, clean, standardize and extract features from the collected data to obtain standardized interaction feature data; S2. Perform node attribute association, edge type matching and weight quantification calculation on standardized interactive feature data to construct a three-dimensional classroom interaction network model of nodes, edges and weights. S3. Based on the real-time dynamic characteristics of the three-dimensional classroom interactive network model, a three-level dynamic evaluation model is constructed and the evaluation is executed. The evaluation score is calculated synchronously, and finally the entire process data is stored and shared with the teaching management system.

Images