A Bayesian Network-Based Method for Evaluating the Effectiveness of Physical Education Teaching

Abstract

This application provides a method for evaluating the effectiveness of physical education teaching based on Bayesian networks, including: calculating the interaction density and intensity distribution of classroom activities per unit time for behavioral feature sequences, extracting key feature indicators, and performing data dimensionality reduction; aligning the dimensionality-reduced classroom behavioral features with standardized physical fitness scores to construct the initial structure of the Bayesian network and determine the relationships between nodes; if there is a lag effect, introducing a time delay factor to adjust the conditional probability distribution in the Bayesian network and capturing the cumulative impact of classroom behavior on physical fitness test results; considering individual differences, calculating individual response coefficients based on students' historical physical fitness test results and classroom behavior data to correct the conditional probability distribution in the Bayesian network; and based on the constructed Bayesian network, inputting real-time classroom behavioral feature data to calculate the probability quantification value of teaching effectiveness and generate teaching effectiveness evaluation results.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for evaluating the effectiveness of physical education teaching based on Bayesian networks. Background Technology

[0002] In the construction and practical research of a training and evaluation system for the dynamic professional development of college physical education teachers based on teaching diagnosis, a key technical issue is how to effectively transform classroom behavior recognition data and student physical fitness test results into probability weight values ​​and establish a quantitative model of teaching effectiveness. Specifically, the classroom behavior recognition system can capture multi-dimensional behavioral data of physical education teachers in the classroom, such as interaction frequency, teaching pace, and guidance methods, while student physical fitness test results reflect students' mastery of motor skills and improvement in physical fitness. These two types of data represent the teaching process and teaching effect, respectively, but their original data forms differ significantly. Classroom behavior data is high-frequency, dynamic time-series data, while physical fitness test results are low-frequency, static numerical data. To unify these two types of data into the same quantitative framework, it is first necessary to extract features and reduce the dimensionality of the classroom behavior data, transforming it into indicators that can match the physical fitness test results. For example, classroom interaction frequency can be transformed into interaction density per unit time, and teaching pace into the intensity distribution of classroom activities. Next, a probabilistic relationship between classroom behavior features and physical fitness test results needs to be constructed using a Bayesian network. This process requires determining the structure and parameters of the Bayesian network, including the conditional probability distribution between nodes, and how to incorporate the impact of different teaching strategies on teaching effectiveness into the model. However, the challenge lies in the fact that the relationship between classroom behavior data and physical fitness test results is not linear and exhibits a lag effect. For example, a high frequency of interaction in a particular class may not be immediately reflected in the physical fitness test results, but rather will only become apparent after multiple lessons have accumulated. Furthermore, different students respond differently to the same teaching strategy, and this heterogeneity needs to be considered in the model. Therefore, designing a quantitative model that can capture the complex relationships between data while simultaneously considering timeliness and individual differences is the core challenge of this technical problem. Summary of the Invention

[0003] This invention provides a method for evaluating the effectiveness of physical education teaching based on Bayesian networks, mainly including:

[0004] Acquire classroom behavior data, including multi-dimensional time-series data on interaction frequency, teaching pace, and guidance methods, and process it in segments through time windows to obtain a sequence of behavioral characteristics per unit time.

[0005] For behavioral feature sequences, calculate the interaction density and intensity distribution of classroom activities per unit time, extract key feature indicators, and complete data dimensionality reduction;

[0006] Obtain students' physical fitness test results, including their mastery of motor skills and improvement in physical fitness, and normalize the test results to obtain standardized physical fitness scores.

[0007] The dimensionality-reduced classroom behavior features are aligned with standardized physical fitness scores to construct the initial structure of a Bayesian network and determine the relationships between nodes.

[0008] Based on historical teaching data, the conditional probability distribution between classroom behavior characteristics and physical fitness test results is calculated, and the parameters of the Bayesian network are determined by maximum likelihood estimation.

[0009] If a lag effect exists, a time delay factor is introduced to adjust the conditional probability distribution in the Bayesian network and capture the cumulative impact of classroom behavior on physical fitness test results.

[0010] To address individual differences, individual response coefficients are calculated based on students' historical physical fitness test results and classroom behavior data, and the conditional probability distribution in the Bayesian network is corrected accordingly.

[0011] Based on the constructed Bayesian network, real-time classroom behavior feature data is input, the probability quantification value of teaching effectiveness is calculated, and teaching effectiveness evaluation results are generated.

[0012] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0013] This invention discloses a method for evaluating the effectiveness of physical education teaching based on Bayesian networks. This method assesses teaching effectiveness by collecting classroom behavior data and student physical fitness test results, and constructing a Bayesian network model. First, the classroom behavior data is segmented into time windows and dimensionality reduced to extract key feature indicators. Then, standardized physical fitness scores are aligned with the behavior feature data to establish the initial structure of the Bayesian network. Conditional probability distributions are calculated using historical data, and adjustments are made considering lag effects and individual differences. Finally, real-time classroom behavior feature data is input to quantitatively evaluate teaching effectiveness. This invention can objectively and dynamically evaluate the effectiveness of physical education teaching, providing teachers with timely feedback, helping to optimize teaching strategies and improve teaching quality. Attached Figure Description

[0014] Figure 1 This is a flowchart of a Bayesian network-based method for evaluating the effectiveness of physical education teaching according to the present invention.

[0015] Figure 2 This is a schematic diagram of a Bayesian network-based method for evaluating the effectiveness of physical education teaching according to the present invention.

[0016] Figure 3 This is another schematic diagram of a Bayesian network-based method for evaluating the effectiveness of physical education teaching according to the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0018] like Figure 1-3 This embodiment of a method for evaluating the effectiveness of physical education teaching based on Bayesian networks may specifically include:

[0019] S101. Obtain classroom behavior data, including multi-dimensional time-series data on interaction frequency, teaching pace, and guidance methods. Process the data in segments through time windows to obtain a sequence of behavioral characteristics per unit time.

[0020] Classroom behavior time-series data is acquired, including three dimensions: interaction count, rhythm value, and mode of instruction. The time-series data is processed using preset segment lengths to obtain multiple time windows. For each time window, feature columns for interaction count, rhythm value, and mode of instruction are extracted. Clustering algorithms are used to classify these feature columns and determine behavioral indicators. If a behavioral indicator belongs to a preset abnormal category, the time window is marked as an abnormal period. A teaching behavior analysis report is generated based on the location and frequency of abnormal periods. A regression model is used to predict future teaching behavior trends within a given time period.

[0021] S102. For the behavioral feature sequence, calculate the interaction density and intensity distribution of classroom activities per unit time, extract key feature indicators, and complete data dimensionality reduction.

[0022] Obtain behavioral feature sequences and calculate interaction density per unit time. Analyze the intensity distribution of classroom activities using the sliding window method. Extract key feature indicators and establish a feature vector space. Perform principal component analysis based on a preset dimensionality reduction threshold. If the feature vector dimension exceeds the threshold, use linear discriminant analysis. Input the dimensionality-reduced feature vectors into a support vector machine to obtain the final behavioral feature classification results.

[0023] S103. Obtain the results of students' physical fitness tests, including the degree of mastery of sports skills and the improvement of physical fitness. Normalize the test results to obtain standardized physical fitness scores.

[0024] Acquire student physical fitness test data, including specific numerical values ​​for the level of mastery of motor skills and the improvement of physical fitness. Calculate skill scores and fitness scores separately for each skill mastery and fitness improvement. Use a normalization method to process the skill and fitness scores, obtaining standardized scores. Generate the final physical fitness assessment result based on the standardized scores. If both the skill and fitness scores are higher than preset thresholds, the physical fitness assessment is judged as excellent. If either the skill or fitness score is lower than the preset threshold, the physical fitness assessment is judged as satisfactory or requires improvement. Compare and analyze the physical fitness assessment results with the original test data from the data source to generate a physical fitness improvement report.

[0025] S104. Align the dimensionality-reduced classroom behavior features with the standardized physical fitness scores to construct the initial structure of the Bayesian network and determine the relationships between nodes.

[0026] Obtain dimensionality-reduced classroom behavior feature data and standardized physical fitness score data. Use a data alignment algorithm to process the dimensionality-reduced and standardized values ​​to obtain aligned feature score data. Construct an initial Bayesian network structure based on the aligned feature score data. Determine the relationships between nodes in the network structure. If strong relationships exist between nodes, add directed edges. Generate a Bayesian network structure graph based on the relationships between nodes. Use a Bayesian network inference algorithm to calculate the conditional probability distribution between nodes.

[0027] ,

[0028] Indicates the number of feature dimensions. Indicates the first Dimensionality reduction values ​​of each feature Indicates the first The mean of each feature, Indicates the first The standard deviation of each feature Indicates the first The weight coefficients of each feature; this formula is used to calculate the score after feature alignment.

[0029] Specifically, processing classroom behavior characteristic data and physical fitness score data involves data dimensionality reduction and alignment. For example, by reducing the dimensionality of multi-dimensional behavioral characteristics such as classroom focus and participation, the original data can be transformed into a more compact representation. Physical fitness scores include two dimensions: motor skills and physical fitness. Standardization makes data from different dimensions comparable. Data alignment algorithms mainly address the scale differences between feature values. Taking classroom performance as an example, a student's focus dimensionality reduction value is 0.8, and their participation dimensionality reduction value is 0.6, corresponding to a physical fitness score of 85. Through alignment, data with different dimensions can be mapped to the same scoring range, making data analysis more reasonable. Constructing a Bayesian network requires determining the relationships between nodes. For example, there may be a causal relationship between classroom focus and motor skill mastery, where students with higher concentration tend to master motor skills more easily. By analyzing data correlations, this relationship can be discovered, and corresponding directed edges can be added to the network. The conditional probability distribution between nodes reflects the quantitative relationship between variables. For example, when a student's classroom focus increases, the probability of mastering motor skills also increases. Specifically, when focus increases from 0.6 to 0.8, the probability of achieving an excellent level of motor skill may increase from 60% to 80%. By setting feature weight coefficients, the importance of different features can be reflected. For example, in assessing basketball skills, shooting accuracy might be given a higher weight, while passing speed might be given a relatively lower weight. This differentiated weighting can more accurately reflect the influence of features on the final score. Data standardization helps eliminate dimensional differences between different features. For instance, converting raw athletic performance into standardized scores can objectively reflect a student's relative position in the overall distribution. When a student's long jump is 180 centimeters, standardization might yield a standardized score of 0.7, making it easier to integrate with other test items for comprehensive evaluation. The results of feature dimensionality reduction and alignment directly affect the accuracy of the Bayesian network. Appropriate data preprocessing can improve the reliability of the network structure, thus providing more valuable references for teaching evaluation. In practical applications, this method can help teachers better understand students' learning patterns and improve teaching strategies accordingly.

[0030] S105. Based on historical teaching data, calculate the conditional probability distribution between classroom behavior characteristics and physical fitness test results, and determine the parameters of the Bayesian network through maximum likelihood estimation.

[0031] Historical teaching data is acquired, and classroom behavior characteristics and physical fitness test results are extracted. Conditional probability distributions are calculated for the extracted classroom behavior characteristics and physical fitness test results. The maximum likelihood estimation method is used to determine the Bayesian network parameters based on the conditional probability distributions. A Bayesian network model is constructed, with classroom behavior characteristics and physical fitness test results as nodes. If the Bayesian network parameters meet a preset threshold, the model is saved; otherwise, the parameter estimation process is readjusted. Based on the Bayesian network model, a relationship graph between classroom behavior characteristics and physical fitness test results is generated. The influence of classroom behavior characteristics on physical fitness test results is analyzed through the relationship graph.

[0032] S106. If a lag effect exists, introduce a time delay factor to adjust the conditional probability distribution in the Bayesian network and capture the cumulative impact of classroom behavior on physical fitness test results.

[0033] Classroom behavior data and physical fitness test data were acquired to construct a time-series dataset. The time-series dataset was input into a pre-established data acquisition module to extract time delay features. A Bayesian network algorithm was employed, using the time delay features as input variables to the network structure. Based on the variable relationships within the network structure, a conditional probability distribution was calculated. If a lag effect exists, a delay factor was introduced to adjust the conditional probability distribution. The cumulative impact of classroom behavior on physical fitness test results was determined using the adjusted conditional probability distribution. A Bayesian network model incorporating the time delay factor was obtained and used to predict physical fitness test results.

[0034] S107. To address individual differences, calculate individual response coefficients based on students' historical physical fitness test results and classroom behavior data, and correct the conditional probability distribution in the Bayesian network.

[0035] This process involves acquiring physical fitness values ​​from historical physical fitness test data and behavioral values ​​from classroom behavior data, and combining them with a pre-defined calculation model to obtain the response value in the individual response coefficient. Based on the response value and the initial conditional probability distribution in the pre-defined Bayesian network, the probability value in the adjusted probability distribution is determined. For physical fitness values, behavioral values, and response values, a multiple regression analysis method is used to calculate the correlation between the individual response coefficient and the physical fitness and behavioral values, resulting in a corrected response value. If the difference between the corrected response value and the initial response value exceeds a pre-defined threshold, the distribution value in the conditional probability distribution of the Bayesian network is updated. Using a Bayesian network inference algorithm, combined with the adjusted probability distribution and the corrected response value, the relationship between individual physical fitness and behavior is inferred, resulting in the network value. Based on the network value and historical values ​​from historical physical fitness test data, the trend of individual physical fitness changes is determined, resulting in the change characteristics in the individual values. Through the change characteristics and classroom behavior values ​​from the classroom behavior data, a cluster analysis method is used to classify student groups, resulting in the group division results in the calculated values.

[0036] S108. Based on the constructed Bayesian network, input real-time classroom behavior feature data, calculate the probability quantification value of teaching effectiveness, and generate teaching effectiveness evaluation results.

[0037] Based on a pre-established Bayesian network model, real-time classroom behavior characteristic data is acquired. If the characteristic data meets a preset threshold, the probability value of teaching effectiveness is calculated using the Bayesian probability calculation method. After determining the calculated probability value, it is checked whether the value is within a preset range. If it is within the range, an effectiveness evaluation report is generated. Based on the content of the evaluation report, corresponding behavioral characteristic parameters are extracted from the network model. If the parameters are abnormal, the probability value of teaching effectiveness is recalculated. Through the extracted characteristic data, iterative calculations are performed in the pre-established Bayesian network model to obtain the final effectiveness evaluation result. Based on the final result, the quantitative value of teaching effectiveness is determined, and it is determined whether there are abnormal behavioral characteristics. If so, the corresponding effectiveness evaluation report is output.

[0038] The above-disclosed embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the invention. Those skilled in the art will understand that implementing all or part of the above-described embodiments and making equivalent changes in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. A method for evaluating the effectiveness of physical education teaching based on Bayesian networks, characterized in that, The method includes: Acquire classroom behavior data, including multi-dimensional time-series data on interaction frequency, teaching pace, and guidance methods, and process it in segments through time windows to obtain a sequence of behavioral characteristics per unit time. For behavioral feature sequences, calculate the interaction density and intensity distribution of classroom activities per unit time, extract key feature indicators, and complete data dimensionality reduction; Obtain students' physical fitness test results, including their mastery of motor skills and improvement in physical fitness, and normalize the test results to obtain standardized physical fitness scores. The dimensionality-reduced classroom behavior features are aligned with standardized physical fitness scores to construct the initial structure of a Bayesian network and determine the relationships between nodes. Based on historical teaching data, the conditional probability distribution between classroom behavior characteristics and physical fitness test results is calculated, and the parameters of the Bayesian network are determined by maximum likelihood estimation. If a lag effect exists, a time delay factor is introduced to adjust the conditional probability distribution in the Bayesian network and capture the cumulative impact of classroom behavior on physical fitness test results. To address individual differences, individual response coefficients are calculated based on students' historical physical fitness test results and classroom behavior data, and the conditional probability distribution in the Bayesian network is corrected accordingly. Based on the constructed Bayesian network, real-time classroom behavior feature data is input, the probability quantification value of teaching effectiveness is calculated, and teaching effectiveness evaluation results are generated. The process involves aligning the dimensionality-reduced classroom behavior features with standardized physical fitness scores, constructing the initial structure of a Bayesian network, and determining the relationships between nodes, including: Obtain dimensionality-reduced classroom behavior characteristic data and standardized physical fitness score data; A data alignment algorithm is used to process the dimensionality reduction and standardization values ​​to obtain aligned feature score data. For the aligned feature scoring data, construct the initial structure of the Bayesian network; Determine the relationships between nodes in the network structure; If there is a strong association between nodes, add directed edges; Generate a Bayesian network structure diagram based on the relationships between nodes; A Bayesian network inference algorithm is used to calculate the conditional probability distribution between nodes; ， Indicates the number of feature dimensions. Indicates the first Dimensionality reduction values ​​of each feature Indicates the first The mean of each feature, Indicates the first The standard deviation of each feature Indicates the first The weight coefficients of each feature; This formula is used to calculate the score after feature alignment; If a lag effect exists, a time delay factor is introduced to adjust the conditional probability distribution in the Bayesian network, capturing the cumulative impact of classroom behavior on physical fitness test results, including: Acquire classroom behavior data and physical fitness test data to construct a time-series dataset; Input the time series dataset into a pre-established data acquisition module to extract time delay features; The Bayesian network algorithm is used, with time delay features as input variables for the network structure; Calculate the conditional probability distribution based on the variable relationships in the network structure; If a lag effect exists, a delay factor is introduced to adjust the conditional probability distribution; The cumulative impact of classroom behavior on physical fitness test results was determined by using the adjusted conditional probability distribution. A Bayesian network model incorporating a time delay factor is obtained to predict physical fitness test results.

2. The method according to claim 1, characterized in that, The acquisition of classroom behavior data includes multi-dimensional time-series data on interaction frequency, teaching pace, and guidance methods. This data is processed through time window segmentation to obtain a sequence of behavioral characteristics per unit time period, including: Acquire time-series data on classroom behavior, including three dimensions: number of interactions, rhythm value, and mode of action; The time series data is processed using a preset segment length to obtain multiple time windows; For each time window, extract feature columns such as interaction count, rhythm value, and mode type; Clustering algorithms are used to classify feature columns and determine behavioral indicators; If the behavior indicator belongs to a preset abnormal category, then mark the time window as an abnormal period. Based on the location and frequency of abnormal periods, a teaching behavior analysis report is generated; A regression model was used to predict future trends in teaching behavior.

3. The method according to claim 1, characterized in that, The process involves calculating the interaction density and intensity distribution of classroom activities per unit time for the behavioral feature sequence, extracting key feature indicators, and performing data dimensionality reduction, including: Obtain behavioral feature sequences and calculate the interaction density per unit time. The sliding window method was used to analyze the intensity distribution of classroom activities; Extract key feature indicators and establish a feature vector space; Principal component analysis is performed based on the preset dimensionality reduction threshold; If the dimension of the feature vector is greater than the threshold, then linear discriminant analysis is used; The reduced feature vectors are then input into the support vector machine. The final behavioral feature classification result is obtained.

4. The method according to claim 1, characterized in that, The process of obtaining student physical fitness test results, including the degree of mastery of motor skills and the improvement of physical fitness, involves normalizing the test results to obtain a standardized physical fitness score, including: Obtain students' physical fitness test data, including specific values ​​on the degree of mastery of sports skills and the improvement of physical fitness; Skill scores and physical fitness scores are calculated separately based on the level of mastery of sports skills and the improvement of physical fitness. The skill score and competency score were processed using a normalization method to obtain standardized score values; The final physical fitness assessment result is generated based on the standardized score. If both the skill score and the quality score are higher than the preset threshold, the physical fitness rating will be excellent. If the skill score or quality score is lower than the preset threshold, the physical fitness assessment is judged as qualified or needs improvement; The physical fitness assessment results are compared and analyzed with the original test data from the data source to generate a physical fitness improvement report.

5. The method according to claim 1, characterized in that, The process of calculating the conditional probability distribution between classroom behavior characteristics and physical fitness test results based on historical teaching data, and determining the parameters of the Bayesian network through maximum likelihood estimation, includes: Acquire historical teaching data and extract classroom behavior characteristics and physical fitness test results; Calculate the conditional probability distribution based on the extracted classroom behavior characteristics and physical fitness test results; The parameters of the Bayesian network are determined by using the maximum likelihood estimation method based on the conditional probability distribution. Construct a Bayesian network model, using classroom behavior characteristics and physical fitness test results as nodes; If the Bayesian network parameters meet the preset threshold, then save the model; If the conditions are not met, the parameter estimation process should be readjusted. Based on the Bayesian network model, a graph showing the relationship between classroom behavior characteristics and physical fitness test results is generated. The influence of classroom behavior characteristics on physical fitness test results was analyzed using a relationship diagram.

6. The method according to claim 1, characterized in that, The method for addressing individual differences involves calculating individual response coefficients based on students' historical physical fitness test results and classroom behavior data, and correcting the conditional probability distribution in the Bayesian network, including: Obtain physical fitness values ​​from students' historical physical fitness test data and behavioral values ​​from classroom behavior data, and combine them with a pre-set calculation model to obtain the response value in the individual response coefficient; Based on the response value and the preset initial conditional probability distribution in the Bayesian network, determine the probability value in the adjusted probability distribution; For physical fitness values, behavioral values, and response values, multiple regression analysis was used to calculate the correlation between individual response coefficients and physical fitness values ​​and behavioral values, and the corrected response values ​​were obtained. If the difference between the corrected response value and the initial response value exceeds a preset threshold, then update the distribution value in the conditional probability distribution in the Bayesian network. Using a Bayesian network inference algorithm, combined with an adjusted probability distribution and corrected response values, the relationship between an individual's physical condition and behavior is inferred, and the network value in the network is obtained. Based on network values ​​and historical values ​​from historical physical fitness test data, determine the trend of individual physical fitness changes and obtain the change characteristics in individual values; By using classroom values ​​from change characteristics and classroom behavior data, cluster analysis is employed to classify student groups, resulting in the calculated group classification results.

7. The method according to claim 1, characterized in that, The constructed Bayesian network, upon inputting real-time classroom behavior characteristic data, calculates the probability quantification value of teaching effectiveness and generates teaching effectiveness evaluation results, including: Based on the pre-established Bayesian network model, real-time classroom behavior feature data is acquired. If the feature data meets the preset threshold, the probability value of teaching effectiveness is calculated using the Bayesian probability calculation method. After determining the calculated probability value, it is judged whether the value is within the preset range. If it is within the range, an effectiveness evaluation report is generated. Based on the evaluation report, relevant behavioral feature parameters are extracted from the network model. If any parameters are abnormal, the probability value of teaching effectiveness is recalculated. The extracted feature data is used to perform iterative calculations in a pre-defined Bayesian network model to obtain the final performance evaluation results. Based on the final results, a quantitative value of teaching effectiveness is determined, and it is determined whether there are any abnormal behavioral characteristics. If so, a corresponding effectiveness evaluation report is output.

Images