Virtual simulation teaching management method and system based on big data

Abstract

The present invention provides a virtual simulation teaching management method and system based on big data, which relates to the field of virtual simulation teaching management technology, including obtaining the behavioral data of students, including the students' operating behaviors, reaction states and emotional reactions, and then scoring the students' behavioral data through concentration analysis. The behavioral data and concentration scores are processed by long short-term memory networks and adaptive fuzzy control algorithms, and the attention fluctuations of students are analyzed and clustered, so as to formulate personalized teaching strategies. Subsequently, by optimizing the students' learning interaction links, and combining the students' concentration intervention strategies with the learning progress data to evaluate the learning efficiency, the teaching progress and path are finally adjusted through a multi-objective optimization algorithm to optimize the teaching management strategy. The method can intelligently adjust the teaching strategies and intervention measures in a data-driven manner, significantly improve the students' learning efficiency and concentration, and contribute to the realization of personalized teaching.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual simulation teaching management, and in particular to a virtual simulation teaching management method and system based on big data. Background Art

[0002] With the rapid development of virtual simulation technology, virtual simulation teaching has become an important means in the field of modern education. Especially in the learning of some complex knowledge, virtual simulation can provide a safer and more flexible learning environment, and help students to practice repeatedly without being restricted by time and space. However, although the existing virtual simulation teaching system has certain teaching effects, the existing technology still has significant problems in improving students' learning efficiency, improving students' concentration and personalized teaching. The current virtual simulation teaching system is usually unable to flexibly adjust the teaching strategy according to the students' real-time behavior and emotional changes, and the teaching content and progress often cannot adapt to the needs of students in a timely manner.

[0003] Therefore, there is an urgent need for a virtual simulation teaching management method and system based on big data to solve the above problems. Summary of the invention

[0004] The purpose of the present invention is to provide a virtual simulation teaching management method and system based on big data to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0005] In a first aspect, the present application provides a virtual simulation teaching management method based on big data, comprising:

[0006] Acquiring behavioral data of a virtual simulation student, wherein the behavioral data includes behavioral data, reaction data, and emotional data of the student;

[0007] Performing a concentration analysis on the behavior data to obtain a concentration score corresponding to the behavior data;

[0008] The behavior data and the concentration score are processed based on a long short-term memory network algorithm and an adaptive fuzzy control algorithm, and cluster analysis is performed based on the processed attention fluctuation data of the students to determine the teaching strategy corresponding to each category of students;

[0009] Optimizing the learning interaction links of the students based on the teaching strategies corresponding to each category of students and the attention fluctuation data of the students to obtain an optimized teaching strategy;

[0010] The optimized teaching strategies and concentration scores are input into the preset learning intervention model for processing to obtain the concentration intervention strategy corresponding to each concentration score;

[0011] The concentration intervention strategy and the preset learning progress data of the students are subjected to efficiency evaluation processing, and the teaching management is optimized based on the learning efficiency evaluation results of the students to obtain an optimized teaching management strategy, which includes teaching progress management and teaching path management.

[0012] In the second aspect, the present application also provides a virtual simulation teaching management system based on big data, including:

[0013] An acquisition unit, used to acquire behavior data of a virtual simulation student, wherein the behavior data includes behavior data, reaction data and emotion data of the student;

[0014] An analysis unit, configured to perform a concentration analysis on the behavior data to obtain a concentration score corresponding to the behavior data;

[0015] A clustering unit, for processing the behavior data and the concentration score based on a long short-term memory network algorithm and an adaptive fuzzy control algorithm, and performing cluster analysis based on the processed attention fluctuation data of the students to determine the teaching strategy corresponding to each category of students;

[0016] An optimization unit, used to optimize the learning interaction link of the students based on the teaching strategy corresponding to each category of students and the attention fluctuation data of the students, so as to obtain an optimized teaching strategy;

[0017] A processing unit, configured to input the optimized teaching strategy and the concentration score into a preset learning intervention model for processing, and obtain a concentration intervention strategy corresponding to each concentration score;

[0018] An evaluation unit is used to perform efficiency evaluation on the concentration intervention strategy and the preset learning progress data of the students, and optimize the teaching management based on the learning efficiency evaluation results of the students to obtain an optimized teaching management strategy, which includes teaching progress management and teaching path management.

[0019] The beneficial effects of the present invention are:

[0020] The present invention obtains the behavior data, reaction data and emotional data of virtual simulation students, combines advanced long short-term memory network algorithms, adaptive fuzzy control algorithms and reinforcement learning models, and analyzes students' concentration, attention fluctuations and emotional states in real time, thereby determining the most suitable personalized teaching strategy for students. By optimizing the students' learning interaction links, improving the students' state adjustment and task switching, the teaching process is made more flexible and efficient. In addition, the application also uses Bayesian reasoning algorithms and multi-objective optimization algorithms to evaluate and optimize the students' learning efficiency in real time, and finally obtains the optimized teaching management strategy, which significantly improves the students' learning effect.

[0021] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 A flowchart of a virtual simulation teaching management method based on big data described in an embodiment of the present invention;

[0024] Figure 2 It is a schematic diagram of the structure of a virtual simulation teaching management system based on big data described in an embodiment of the present invention.

[0025] In the figure: 701, acquisition unit; 702, analysis unit; 703, clustering unit; 704, optimization unit; 705, processing unit; 706, evaluation unit. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0027] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0028] Embodiment 1:

[0029] This embodiment provides a virtual simulation teaching management method based on big data.

[0030] See also Figure 1 , the figure shows that the method includes step S1, step S2, step S3, step S4, step S5 and step S6.

[0031] Step S1, obtaining behavior data of a virtual simulation student, wherein the behavior data includes behavior data, reaction data and emotion data of the student;

[0032] It is understandable that the data collection in this step mainly relies on real-time monitoring by sensors, monitoring equipment and intelligent algorithms in the virtual simulation system. Behavioral data mainly reflects the operation process of students in the virtual simulation environment, such as mouse clicks, keyboard input, operation time, reaction speed, etc.; reaction data focuses on the speed and accuracy of students' response to tasks and challenges, and often measures students' physiological changes (such as heart rate, skin electrical response, etc.) through sensors; emotional data analyzes students' emotional state through facial recognition technology, voice recognition technology, etc., and further reflects their emotional fluctuations, such as anxiety, fatigue or active participation. The acquisition of these data can not only comprehensively record students' behavior, but also provide a multi-dimensional basis for subsequent learning status analysis and concentration assessment.

[0033] Step S2, performing concentration analysis on the behavior data to obtain a concentration score corresponding to the behavior data;

[0034] It can be understood that this step accurately analyzes the students' concentration in a digital and quantitative way, avoiding the subjective evaluation or indirect speculation of students' concentration in traditional teaching. The acquisition of concentration scores can reflect the fluctuations of students' concentration in different learning tasks, help the teaching system to evaluate the students' learning status in real time, and provide data support for subsequent personalized teaching adjustments and intervention measures. In this step, step S2 includes step S21, step S22, step S23, step S24 and step S25.

[0035] Step S21: Analyze the time series of the students' behavior data based on the dynamic time warping algorithm to obtain the change data of the students' operation reaction speed and accuracy, and obtain the operation fluency score of each student by calculating the dynamic time warping distance, and use it as the concentration score corresponding to the behavior data, wherein the behavior data includes operation time data, reaction speed data and error rate data;

[0036] It is understandable that this step uses the dynamic time warping algorithm to accurately capture the subtle differences in the students' task execution, thereby providing a more reliable basis for concentration scoring. Unlike traditional static analysis methods, dynamic time warping can handle nonlinear and time-distorted behavior patterns, allowing students' performance to be evaluated more realistically and accurately.

[0037] In this step, the calculation formula of the dynamic time warping distance is as follows:

[0038] in, represents the dynamic time warping distance between two data, and Represents two different data in The value at a time point, Indicates that there is a point in time.

[0039] Step S22, performing multi-scale analysis on the reaction data by wavelet transform, wherein the reaction state of the trainee is obtained by extracting high-frequency and low-frequency signal change data in the reaction data, and a concentration score corresponding to the reaction data is determined based on a preset score threshold corresponding to the reaction state, wherein the reaction data includes heart rate data and skin electrical response data;

[0040] It is understood that the response data in this step usually includes the heart rate and skin galvanic response data of the trainee, which can be collected in real time by wearing sensors or monitoring equipment. The heart rate response is usually closely related to the trainee's emotional fluctuations and changes in concentration, while the skin galvanic response can reflect the trainee's physiological stress response, which is usually related to the trainee's anxiety, stress or concentration.

[0041] It can be understood that in this step, the high concentration state usually corresponds to a heart rate range of 60 to 75 beats / minute, and the galvanic skin response is between 0.1 and 0.3 microSiemens. This type of state is assigned a concentration score between 8 and 10. The heart rate range for the medium concentration state is 76 to 90 beats / minute, the galvanic skin response is 0.4 to 0.6 microSiemens, and the scoring threshold is set between 5 and 7. For the low concentration state, the heart rate is usually between 91 and 110 beats / minute, the galvanic skin response range is 0.7 to 1.0 microSiemens, and the scoring threshold is between 3 and 4. The distracted state is manifested as a heart rate exceeding 110 beats / minute, a galvanic skin response exceeding 1.0 microSiemens, and the corresponding concentration scoring threshold is 1 to 2. These scoring thresholds are not fixed ranges, but intervals obtained through a large amount of statistical analysis of data, which are used to quantify the concentration levels of different reaction states.

[0042] Before wavelet transform, these reaction data are preprocessed to remove noise and unnecessary interference. Usually, the preprocessing steps include data filtering and normalization. Through wavelet transform, the reaction data is decomposed into different scale (frequency) components, including high-frequency components and low-frequency components. Among them, high-frequency components usually represent the fast-changing parts of the data, reflecting the fluctuations of the students' reactions in a short period of time. These fluctuations may be related to factors such as the students' emotional changes, anxiety or tension. Low-frequency components usually represent the slowly changing parts of the data, reflecting the long-term trend or stable state of the students, such as a continuous state of concentration or stable emotional fluctuations. By analyzing the changes in the extracted high-frequency and low-frequency signals, the reaction state of the students can be obtained.

[0043] Step S23, extracting features of emotion data by using an adaptive convolutional neural network, wherein the emotion data includes a facial image and an audio signal of the trainee, and extracting detailed features of the trainee's facial muscle movements and time-frequency features of the audio signal of the trainee's communication by using the adaptive convolutional neural network;

[0044] It is understandable that this step can automatically extract features from the student's facial image through an adaptive convolutional neural network. The convolutional neural network can extract the movement features of facial muscles such as eyebrows and mouth corners from the original facial image through multiple convolutional layers, pooling layers and other structures, and then finally obtain the detailed features of facial muscle movements that can reflect the student's emotional changes through the deep structure of the network. In this step, the audio signal is the student's voice data collected through a microphone, the audio signal is converted into a time-frequency domain representation, the frequency components of the signal in different time periods are extracted, and the frequency components of the signal in different time periods are input into the adaptive convolutional neural network for analysis, and the emotional information in the voice is identified, including the rise and fall of the pitch and the change of the speaking speed, and then the voice characteristics of the student are output.

[0045] Step S24, combining features based on the detailed features of the facial muscle movements of the trainee and the time-frequency features of the audio signal of the trainee's communication, and determining a concentration score corresponding to the emotional state based on the emotional state of the trainee obtained by combining and a score threshold corresponding to the preset emotional state;

[0046] It can be understood that this step assigns different weights to signals of different modes, reflecting their relative importance to emotional judgment, and obtains comprehensive characteristics of emotional state. By combining the time-frequency characteristics of facial muscle movements and audio signals, this step can effectively improve the accuracy of emotional state assessment. Compared with a single data source, the fusion of multimodal data helps to avoid the bias caused by a certain signal source, thereby providing more comprehensive emotion recognition. The introduction of adaptive convolutional neural networks further enhances the system's ability to adapt to individual differences, allowing students' emotional changes to be accurately captured and evaluated.

[0047] It is understandable that when assessing the emotional state of trainees, facial muscle movement characteristics and speech time-frequency characteristics are key indicators. The calm state is usually characterized by facial muscle movement changes of less than 15%, speech time-frequency changes of less than 100 Hz, and the corresponding emotional score threshold is 8 to 10. In a slightly anxious state, facial muscle movement changes between 16% and 30%, speech time-frequency changes between 101 Hz and 200 Hz, and the score threshold is set at 5 to 7. In an anxious state, facial muscle movement changes between 31% and 50%, speech time-frequency changes between 201 Hz and 300 Hz, and the emotional score threshold is 3 to 4. A highly anxious state is characterized by facial muscle movement changes of more than 50%, speech time-frequency changes of more than 300 Hz, and the corresponding emotional score threshold is 1 to 2. These scoring thresholds are derived from the analysis of a large amount of data. They are not fixed ranges, but tools for quantifying and distinguishing different emotional states, helping to more accurately assess trainees' emotional responses.

[0048] Step S25: Perform weighted calculation on the concentration score corresponding to the behavior data, the concentration score corresponding to the reaction data, and the concentration score corresponding to the emotional state to obtain the concentration score corresponding to the behavior data.

[0049] It is understandable that in this step, through weighted calculation, the system can integrate multi-dimensional data to more accurately evaluate the student's concentration. This weighted method not only takes into account the relative importance of different data sources, but also dynamically adjusts the weights according to the actual application scenarios, thereby achieving a personalized assessment of the student's concentration. For example, if a student has large emotional fluctuations and strong physiological reactions, the system may give priority to emotional and physiological data, thereby increasing the sensitivity of the concentration score. For a student whose behavioral data performs well, the weight of the behavioral data may be increased to better reflect the student's actual learning concentration.

[0050] Step S3, processing the behavior data and the concentration score based on the long short-term memory network algorithm and the adaptive fuzzy control algorithm, and performing cluster analysis based on the processed attention fluctuation data of the students to determine the teaching strategy corresponding to each category of students;

[0051] It can be understood that this step combines the advantages of the long short-term memory network algorithm and the adaptive fuzzy control algorithm, so that the student's attention fluctuation data can be accurately modeled and dynamically adjusted. This data processing method can better reflect the real-time status of the students than the traditional method, and can adjust the prediction strategy in time during the actual teaching process, thereby improving the adaptability and accuracy of teaching management. Through cluster analysis, the system can classify students according to their concentration characteristics, thereby providing personalized and targeted teaching strategies. Compared with the traditional "one-size-fits-all" teaching method, this method can provide more effective teaching interventions according to the different needs of students, greatly improving students' learning efficiency and learning experience.

[0052] It can be understood that step S3 in this step includes step S31, step S32 and step S33.

[0053] Step S31, arranging the behavior data and the concentration score into time series data in chronological order, and inputting the time series data into a long short-term memory network for training and prediction, to obtain the student's attention fluctuation prediction data;

[0054] It can be understood that this step inputs the student's behavior data (such as operation time, accuracy, etc.) and concentration scores into the long short-term memory network in chronological order. Through the training of the long short-term memory network, the student's behavior pattern and concentration change trend are extracted, wherein the long short-term memory network controls the information flow through the internal gating mechanism (such as input gate, forget gate and output gate), so as to capture the long-term dependence of the student's attention fluctuation during the learning process. During the training process, the long short-term memory will learn how to predict the student's future concentration fluctuation based on the student's historical data (such as past behavior data and concentration scores). The long short-term memory model predicts the student's attention fluctuation data for a period of time in the future based on the existing data, and outputs the student's attention fluctuation trend in the future. For example, the system can predict what attention fluctuations the student may experience in the next period of time and give the corresponding concentration score. These prediction results can help teachers understand the student's future learning status and adjust the teaching strategy.

[0055] Step S32, using the attention fluctuation prediction data as input to an adaptive fuzzy control algorithm, wherein the adaptive fuzzy control algorithm adjusts the control rules in the adaptive fuzzy control algorithm by comparing the actual concentration score with the prediction result, and then obtains the prediction result of the student's concentration score through a fuzzy reasoning mechanism;

[0056] It can be understood that this step calculates the error between the predicted result and the actual concentration score. The error value is usually expressed as the difference between the actual value and the predicted value, which may be positive or negative, indicating that the prediction is too high or too low. According to the size and direction of the error, the rules in the fuzzy control system are automatically adjusted. For example, if the predicted concentration is too low, the system will adjust the relevant rules so that the next prediction result can be closer to the actual concentration. The adjusted control rules are applied to the fuzzy reasoning system to re-predict the student's concentration. This step uses the preset fuzzy rule base and the fuzzy reasoning mechanism to infer the input fuzzified data. For example, "when the concentration fluctuates greatly and the predicted value is low, increase the concentration intervention."

[0057] Step S33: Based on the K-means clustering algorithm and the prediction results of the concentration score, the students are classified into groups, and the mean score of each category of students is subjected to grey correlation analysis with the historical teaching strategy to obtain the teaching strategy corresponding to each category of students.

[0058] It can be understood that this step uses the predicted data of the fluctuation of the students' concentration obtained previously to organize the students' concentration scores and fluctuation trends according to the time series into feature vectors. These vectors serve as input data for K-means clustering. Then select an appropriate number of clusters K (that is, how many categories the students are divided into). Among them, the selection of the number of clusters K can be determined by the silhouette coefficient. This decision is based on the characteristic distribution of the student group to ensure that the grouping results have good discrimination. In this step, the K-means algorithm groups the students according to the initial cluster center (mean) in an iterative manner. In each iteration, the algorithm assigns each student to the nearest cluster center and updates the position of the cluster center until the cluster center no longer changes or changes very little, indicating that the clustering process has converged.

[0059] The historical teaching strategy data in this step refers to the relationship between the previously implemented teaching strategies and the learning outcomes of the students. These strategies include teaching methods, selection of learning materials, design of interactive tasks, etc. The grey correlation analysis in this step is used to quantify the similarity between the student groups and different teaching strategies. By calculating the similarity (i.e. grey correlation) between the mean characteristics of the student group and each historical teaching strategy, it is possible to determine which teaching strategies are most suitable for a certain student group. By calculating the difference between the mean of each group (such as the mean of the concentration score) and the historical strategy, a correlation matrix is ​​obtained, which represents the adaptability of different teaching strategies to different student groups. In this step, based on the results of the grey correlation, the most effective teaching strategy for each student group can be selected. For example, if a student group performs better under a strategy with a higher grey correlation of the historical strategy, then this strategy will be selected as the teaching strategy for that group in the future.

[0060] This step uses the K-means clustering algorithm to classify students into groups, and uses grey correlation analysis to evaluate the correlation between each student group and the history teaching strategy. This step provides accurate data support for teaching management. In this way, personalized teaching strategies can be formulated for each student group, thereby improving the adaptability and effectiveness of teaching, and maximizing students' learning effects and concentration in practical applications.

[0061] Step S4, optimizing the learning interaction link of the students based on the teaching strategy corresponding to each category of students and the attention fluctuation data of the students to obtain an optimized teaching strategy;

[0062] It can be understood that this step optimizes the learning interaction link based on the student's attention fluctuation data and the teaching strategy of each category, and this step effectively improves the student's learning concentration and learning effect. In this step, step S4 includes step S41, step S42 and step S43.

[0063] Step S41: According to the teaching strategy corresponding to each category of students and the students' attention fluctuation information, a reinforcement learning algorithm is used to optimize the students' learning interaction process to obtain the students' state adjustment strategy;

[0064] It can be understood that this step is first modeled by the reinforcement learning algorithm. In this step, the learning state of the students mainly includes their concentration, emotional state, task execution progress, etc., which jointly determine the current learning state of the students. The student's attention fluctuation information is the key input, reflecting the student's concentration during the learning process. Among them, the input (state space) of the reinforcement learning model includes the student's concentration fluctuation data, emotional data (such as facial expressions and physiological reactions), and the difficulty and progress of the task. These input data will constitute the student's "state" and reflect the student's performance in all aspects of the learning process. The "action" of the reinforcement learning algorithm represents the adjustment of the teaching strategy. That is, under the current student state, the teaching adjustment measures that the system can take may include changing the difficulty of the task, adjusting the interactive form of the task, controlling the duration of the task, or guiding the student to switch between different types of tasks. Then, according to the fluctuation of the student's concentration, set the reward mechanism. For example, when the student maintains a high degree of concentration, the system will give positive rewards; on the contrary, if the student has a decrease in attention, low mood, etc., a negative reward will be given. This mechanism encourages students to stay focused and avoid inefficient learning due to distraction.

[0065] Through repeated learning and trial and error, the reinforcement learning model continuously updates and optimizes its strategy. Each interactive feedback from the students (such as changes in concentration, emotional reactions, etc.) provides new training data for the reinforcement learning model, enabling it to make more accurate adjustments to the teaching strategy in the next round of interaction. The students' attention fluctuations and emotional states are constantly changing. The design and difficulty of the tasks in this step should be adjusted as the students' states change. For example, for students with large fluctuations in attention, the system automatically reduces the complexity of the task or provides frequent feedback to help students refocus.

[0066] The reward function in the reward mechanism in this step is as follows:

[0067] in, represents the reward function, , , Represents the weight coefficient in the reward function, which is used to balance the influence of concentration, emotional state and concentration fluctuation. and To adjust the weight coefficient of the reward range and control the difference in rewards for high and low concentration, and is the coefficient that adjusts the reward amplitude of the emotional state, For students at time step The concentration score of is the concentration fluctuation amplitude, that is, the concentration difference between the current time step and the previous time step, For students at time step The emotional state score, is the emotion threshold, For students at time step Focus score.

[0068] Step S42: analyzing the switching frequency and task complexity of the trainee between different tasks according to the trainee's state adjustment strategy and task switching theory, and adjusting the interactive task based on the switching frequency and task complexity between the different tasks to obtain an adjusted interactive task;

[0069] It is understandable that in this step, when the student's concentration is high, the frequency of task switching can be increased to give the student more challenges to maintain his or her high concentration; when the student's concentration fluctuates or decreases, the system can reduce the frequency of task switching to give the student more time to concentrate on completing the current task and avoid excessive burden. The student's state adjustment strategy will optimize the complexity of the task. For example, if the student's emotional state is poor or the concentration is low, the system can automatically reduce the difficulty of the task or simplify the content of the task to help the student regain his or her attention; when the student is in a better state, the system can appropriately increase the difficulty of the task and increase its cognitive load to stimulate the student's learning potential.

[0070] Step S43: performing iterative genetic operations on the adjusted interactive tasks based on a genetic algorithm, wherein the optimized interactive tasks corresponding to each student are obtained through iterative genetic operations, and the teaching strategies corresponding to each category of students and the optimized interactive tasks corresponding to each student are used as the optimized teaching strategies.

[0071] It can be understood that in this step, the genetic algorithm first selects a part of the best performing "individuals" from the "population". These "individuals" are the interactive task plans of the students. The effect and fitness of the tasks will be evaluated according to the preset fitness function. The fitness function should be calculated based on indicators such as the change in the students' concentration, emotional fluctuations, and task completion during the interaction process. For example, if the students show a high degree of concentration during the interactive task (derived from the concentration score data in the previous step) and can successfully complete the task, the fitness of the task plan is high; on the contrary, if the students are distracted or have large emotional fluctuations, the fitness of the task plan is low.

[0072] The fitness function is as follows:

[0073] in, For fitness score, Rate participants' focus to measure how well they stay focused on the task. The frequency of task switching of the trainees was measured by the number of times the trainees switched between tasks. It is the students’ task completion degree, reflecting the quality of students’ completion of the task. For students’ emotional stability, It indicates the amplitude of students' emotional fluctuations during the task. It is the complexity of the task, indicating the difficulty of the task itself. , , , and They are adjustment coefficients used to balance the importance of various factors.

[0074] After the selection operation, the genetic algorithm will perform "crossover" and "mutation" operations. The crossover operation is equivalent to combining the task plans of the two students to generate a new task plan. After the crossover, some more innovative task arrangements may be generated. This is because through the crossover operation, the system can combine the successful experience and optimization strategies of the two students to further improve the effect of the learning task. The mutation operation is to add randomness within a certain range to avoid falling into the local optimal solution during the search process and ensure the diversity of tasks. During processing, mutation can make the task more personalized by adjusting parameters such as the difficulty of the task and the switching frequency between tasks, which is in line with the cognitive level and attention changes of different students.

[0075] Through multiple generations of iterations, the genetic algorithm continuously optimizes the interactive tasks of the students. The calculation of the "fitness" of each generation will depend on the actual performance data of the students, including fluctuations in concentration, mood swings, task completion, etc. The result of each iteration is the basis for the optimization task plan of the previous generation until the optimal task configuration is achieved. In order to improve efficiency, certain stopping criteria can also be set. For example, if the fitness reaches stability within a certain number of generations, the iteration can be stopped.

[0076] Step S5: inputting the optimized teaching strategy and concentration score into a preset learning intervention model for processing to obtain a concentration intervention strategy corresponding to each concentration score;

[0077] It is understandable that this step can dynamically adjust the teaching strategy according to the student's concentration score, ensure that the teaching content matches the student's learning status, and enhance the learning experience. The adaptive ability of the reinforcement learning model enables each student to get the most appropriate learning support according to their needs. In a virtual simulation teaching environment, the student's concentration fluctuates greatly, especially in complex operation tasks and long-term learning processes. By inputting the optimized teaching strategy and concentration score into the reinforcement learning model, the system can automatically adjust the teaching strategy according to the student's concentration. For example, in a skill training task, if the system detects that the student's attention is reduced, the reinforcement learning model will arrange a short break to help the student restore his concentration and improve learning efficiency. In this step, step S5 includes step S51 and step S52.

[0078] Step S51: using the Bayesian inference algorithm to perform state inference based on the student's concentration score and behavior data, wherein the student's current concentration state is predicted by updating the posterior probability distribution of the student's concentration;

[0079] It can be understood that in this step, based on the updated posterior probability distribution, the system can predict the student's current concentration state. This prediction is based on all observed behavioral data, so it can more accurately reflect the student's actual learning state. The predicted concentration state will serve as the basis for the next teaching decision. For example, if the student's concentration is predicted to be low, the system may make a concentration intervention strategy to restore concentration.

[0080] Step S52: Send the information to a preset learning intervention model based on the student's current concentration state, and formulate an intervention strategy for the student's rest in the concentration state, wherein the intervention strategy includes the student's rest time.

[0081] It is understandable that in order to formulate an appropriate rest intervention strategy in this step, it is necessary to first define the threshold of the concentration state. For example: High concentration state: A high concentration score indicates that the student is in a good state of attention and may not need to rest immediately. Medium concentration state: The concentration score is at a medium level, and the student may be on the verge of attention fatigue and may need a short break. Low concentration state: A low concentration score indicates that the student's attention has decreased significantly, and the system should recommend that the student take a longer break to restore energy.

[0082] Depending on the student's concentration level, the intervention model can adjust the break time. Common break strategies include: Short break: For students with high or medium concentration, the model may recommend a short break of 5 to 10 minutes. Long break: For students with low concentration, a longer break, such as 15 to 20 minutes, may be needed to fully restore concentration.

[0083] In addition to the duration of breaks, the learning intervention model can also determine the frequency of breaks. For example, for students who study continuously for a long time, the system may prompt them to schedule a break every 45 minutes to 1 hour.

[0084] Step S6: perform efficiency evaluation on the concentration intervention strategy and the preset student's learning progress data, and optimize the teaching management based on the student's learning efficiency evaluation results to obtain an optimized teaching management strategy, which includes teaching progress management and teaching path management.

[0085] It is understandable that this step optimizes the teaching management strategy and ensures the intelligent adjustment of the teaching progress and path by combining the concentration intervention strategy with the student's learning progress data and performing efficiency evaluation. This process can not only improve the learning efficiency of students, but also improve the flexibility and effectiveness of the entire teaching system on the basis of personalized management, so that the virtual simulation teaching environment can better meet the actual needs of students and improve teaching results. In this step, step S6 includes step S61 and step S62.

[0086] Step S61: construct a learning efficiency evaluation model based on the concentration intervention strategy and the preset learning progress data of the students, and calculate the learning efficiency of each student under the concentration intervention strategy;

[0087] It can be understood that in this step, the system builds a learning efficiency evaluation model based on the student's learning progress data and concentration intervention strategy. The model generally includes the following main parameters:

[0088] Task completion rate: measures the proportion of tasks completed by trainees within the specified time, reflecting the trainees' task execution status.

[0089] Time efficiency: Compare the time it takes for students to complete the task with the standard time. Too long a completion time may mean that the student is not focused enough or the task is too difficult.

[0090] Task quality: The quality of students’ task completion is evaluated through indicators such as error rate and accuracy, reflecting the relationship between concentration and task execution quality.

[0091] Emotional state: By monitoring students’ emotional data (such as facial expressions, voice emotion analysis, etc.), the impact of emotional fluctuations on learning efficiency can be evaluated.

[0092] Among them, by weighted calculation of the task completion score, time efficiency score, task quality score and emotional state score, the learning efficiency of each student under the concentration intervention strategy was obtained.

[0093] Step S62: Based on the multi-objective optimization algorithm, multi-objective optimization of the teaching management strategy is performed on the learning efficiency and the preset test scores of the students to obtain an optimized teaching management strategy.

[0094] It can be understood that the multi-objective optimization problem in this step involves at least two goals: learning efficiency and test scores. Learning efficiency is usually measured by factors such as task completion, time efficiency, task quality, and emotional state, while test scores measure students' performance at the end of the semester or in periodic tests. The goal of multi-objective optimization is to improve these two indicators at the same time, avoiding excessive bias towards one goal and neglecting the other. The multi-objective optimization algorithm comprehensively considers the students' learning efficiency and test scores, and optimizes the teaching management strategy. This method can not only balance the relationship between efficiency and grades in the students' learning process, but also realize personalized teaching strategies and dynamic adjustments to ensure the maximization of teaching effectiveness. Through multi-objective optimization, the education system can accurately locate the best learning path and intervention strategy for each student, thereby improving the overall teaching quality and the students' learning experience. Among them, the multi-objective optimization function in this step is as follows:

[0095] in, represents the comprehensive optimization function, Represents the weight coefficient of learning efficiency and test scores, Represents the global weight coefficient, which controls the priority of the test scores. represents the weight of task completion, represents the weight of time efficiency, represents the weight of the task quality, The weight representing the level of concentration, Indicates the number of completed tasks. Indicates the total number of tasks, Indicates the actual time spent by students. Indicates the best time, represents the task quality score, Indicates the student's concentration score. Indicates the test scores of the students.

[0096] Embodiment 2:

[0097] like Figure 2 As shown, this embodiment provides a virtual simulation teaching management system based on big data, see Figure 2 The system includes an acquisition unit 701 , an analysis unit 702 , a clustering unit 703 , an optimization unit 704 , a processing unit 705 and an evaluation unit 706 .

[0098] An acquisition unit 701 is used to acquire behavior data of a virtual simulation student, wherein the behavior data includes behavior data, reaction data and emotion data of the student;

[0099] An analysis unit 702 is used to perform a concentration analysis on the behavior data to obtain a concentration score corresponding to the behavior data;

[0100] A clustering unit 703 is used to process the behavior data and the concentration score based on a long short-term memory network algorithm and an adaptive fuzzy control algorithm, and perform cluster analysis based on the processed attention fluctuation data of the students to determine the teaching strategy corresponding to each category of students;

[0101] The optimization unit 704 is used to optimize the learning interaction link of the students based on the teaching strategy corresponding to each category of students and the attention fluctuation data of the students to obtain an optimized teaching strategy;

[0102] The processing unit 705 is used to input the optimized teaching strategy and the concentration score into a preset learning intervention model for processing to obtain a concentration intervention strategy corresponding to each concentration score;

[0103] The evaluation unit 706 is used to perform efficiency evaluation on the concentration intervention strategy and the preset student's learning progress data, and optimize the teaching management based on the student's learning efficiency evaluation results to obtain an optimized teaching management strategy, which includes teaching progress management and teaching path management.

[0104] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention. The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A virtual simulation teaching management method based on big data, characterized in that: include: Acquiring behavioral data of a virtual simulation student, wherein the behavioral data includes behavioral data, reaction data, and emotional data of the student; Performing a concentration analysis on the behavior data to obtain a concentration score corresponding to the behavior data; The behavior data and the concentration score are processed based on a long short-term memory network algorithm and an adaptive fuzzy control algorithm, and cluster analysis is performed based on the processed attention fluctuation data of the students to determine the teaching strategy corresponding to each category of students; Optimizing the learning interaction links of the students based on the teaching strategies corresponding to each category of students and the attention fluctuation data of the students to obtain an optimized teaching strategy; The optimized teaching strategies and concentration scores are input into the preset learning intervention model for processing to obtain the concentration intervention strategy corresponding to each concentration score; The concentration intervention strategy and the preset learning progress data of the students are subjected to efficiency evaluation processing, and the teaching management is optimized based on the learning efficiency evaluation results of the students obtained by the evaluation, so as to obtain an optimized teaching management strategy, wherein the teaching management strategy includes teaching progress management and teaching path management; The behavior data and the concentration score are processed based on the long short-term memory network algorithm and the adaptive fuzzy control algorithm, and cluster analysis is performed based on the processed attention fluctuation data of the students to determine the teaching strategy corresponding to each category of students, including: Arrange the behavior data and the concentration score into time series data in chronological order, and input the time series data into a long short-term memory network for training and prediction, so as to obtain the student's attention fluctuation prediction data; The attention fluctuation prediction data is used as the input of the adaptive fuzzy control algorithm, wherein the adaptive fuzzy control algorithm adjusts the control rules in the adaptive fuzzy control algorithm by comparing the actual concentration score with the prediction result, and then obtains the prediction result of the student's concentration score through the fuzzy reasoning mechanism; Based on the K-means clustering algorithm and the prediction results of concentration scores, students are classified into groups, and the mean score of each category of students is subjected to grey correlation analysis with the historical teaching strategy to obtain the corresponding teaching strategy for each category of students.

2. The virtual simulation teaching management method based on big data according to claim 1 is characterized in that , performing a concentration analysis on the behavior data to obtain a concentration score corresponding to the behavior data, including: Based on the dynamic time warping algorithm, the time series analysis of the students' behavioral data is carried out to obtain the change data of the students' operation reaction speed and accuracy, and the operation fluency score of each student is obtained by calculating the dynamic time warping distance, which is used as the concentration score corresponding to the behavioral data, wherein the behavioral data includes operation time data, reaction speed data and error rate data; Performing multi-scale analysis on the reaction data by wavelet transform, wherein the reaction state of the trainee is obtained by extracting high-frequency and low-frequency signal change data in the reaction data, and determining the concentration score corresponding to the reaction data based on a preset score threshold corresponding to the reaction state, wherein the reaction data includes heart rate data and skin electrical response data; Extracting features of emotion data by using an adaptive convolutional neural network, wherein the emotion data includes a facial image and an audio signal of the trainee, and extracting detailed features of the trainee's facial muscle movements and time-frequency features of the audio signal of the trainee's communication by using the adaptive convolutional neural network; Combining features based on the detailed features of the facial muscle movements of the trainee and the time-frequency features of the audio signal of the trainee's communication, and determining a concentration score corresponding to the emotional state based on the emotional state of the trainee obtained by combining and a score threshold corresponding to the preset emotional state; The concentration score corresponding to the behavior data, the concentration score corresponding to the reaction data, and the concentration score corresponding to the emotional state are weightedly calculated to obtain the concentration score corresponding to the behavior data.

3. The virtual simulation teaching management method based on big data according to claim 1 is characterized in that , based on the teaching strategies corresponding to each category of students and the attention fluctuation data of the students, the learning interaction links of the students are optimized, including: According to the teaching strategies corresponding to each category of students and the students' attention fluctuation information, the reinforcement learning algorithm is used to optimize the students' learning interaction process and obtain the students' state adjustment strategy; Analyzing the switching frequency and task complexity of the trainee between different tasks according to the trainee's state adjustment strategy and task switching theory, and adjusting the interactive task based on the switching frequency and task complexity between the different tasks to obtain an adjusted interactive task; Based on the genetic algorithm, iterative genetic operations are performed on the adjusted interactive tasks, wherein the optimized interactive tasks corresponding to each student are obtained through iterative genetic operations, and the teaching strategies corresponding to each category of students and the optimized interactive tasks corresponding to each student are used as the optimized teaching strategies.

4. The virtual simulation teaching management method based on big data according to claim 1 is characterized in that ,Based on the optimized teaching strategy and concentration score, it is input into the preset learning intervention model for processing, including: Based on the students' concentration scores and behavior data, the Bayesian inference algorithm is used to infer the state. The posterior probability distribution of the students' concentration is updated to predict the students' current concentration state. According to the student's current concentration state, it is sent to the preset learning intervention model to formulate an intervention strategy for the student's rest in the concentration state, wherein the intervention strategy includes the student's rest time.

5. A virtual simulation teaching management system based on big data, characterized in that: include: An acquisition unit, used to acquire behavior data of a virtual simulation student, wherein the behavior data includes behavior data, reaction data and emotion data of the student; An analysis unit, configured to perform a concentration analysis on the behavior data to obtain a concentration score corresponding to the behavior data; A clustering unit, for processing the behavior data and the concentration score based on a long short-term memory network algorithm and an adaptive fuzzy control algorithm, and performing cluster analysis based on the processed attention fluctuation data of the students to determine the teaching strategy corresponding to each category of students; An optimization unit, used to optimize the learning interaction link of the students based on the teaching strategy corresponding to each category of students and the attention fluctuation data of the students, so as to obtain an optimized teaching strategy; A processing unit, configured to input the optimized teaching strategy and the concentration score into a preset learning intervention model for processing, and obtain a concentration intervention strategy corresponding to each concentration score; An evaluation unit, configured to perform efficiency evaluation on the concentration intervention strategy and the preset learning progress data of the students, and optimize the teaching management based on the learning efficiency evaluation results of the students, to obtain an optimized teaching management strategy, wherein the teaching management strategy includes teaching progress management and teaching path management; Wherein, the clustering unit comprises: A first clustering subunit is used to organize the behavior data and the concentration score into time series data in chronological order, and input the time series data into a long short-term memory network for training and prediction, so as to obtain the student's attention fluctuation prediction data; A second clustering subunit is used to use the attention fluctuation prediction data as an input of an adaptive fuzzy control algorithm, wherein the adaptive fuzzy control algorithm adjusts the control rules in the adaptive fuzzy control algorithm by comparing the actual concentration score with the prediction result, and then obtains the prediction result of the student's concentration score through a fuzzy reasoning mechanism; The third clustering subunit is used to classify students into groups based on the prediction results of the K-means clustering algorithm and the concentration score, and to perform grey correlation analysis on the mean score of each category of students and the historical teaching strategy to obtain the corresponding teaching strategy for each category of students.

6. The virtual simulation teaching management system based on big data according to claim 5 is characterized in that: The analysis unit comprises: The first analysis subunit is used to analyze the time series of the students' behavior data based on the dynamic time warping algorithm to obtain the change data of the students' operation reaction speed and accuracy, and obtain the operation fluency score of each student by calculating the dynamic time warping distance, which is used as the concentration score corresponding to the behavior data, wherein the behavior data includes operation time data, reaction speed data and error rate data; A second analysis subunit is used to perform multi-scale analysis on the reaction data by wavelet transform, wherein the reaction state of the trainee is obtained by extracting high-frequency and low-frequency signal change data in the reaction data, and a concentration score corresponding to the reaction data is determined based on a preset score threshold corresponding to the reaction state, wherein the reaction data includes heart rate data and skin electrical response data; The third analysis subunit is used to extract features of emotion data through an adaptive convolutional neural network, wherein the emotion data includes a facial image and an audio signal of the trainee, and the adaptive convolutional neural network is used to extract detailed features of the trainee's facial muscle movements and time-frequency features of the audio signal of the trainee's communication; A fourth analysis subunit is used to combine features based on the detailed features of the facial muscle movements of the trainee and the time-frequency features of the audio signal of the trainee's communication, and determine a concentration score corresponding to the emotional state based on the emotional state of the trainee obtained by combining and a score threshold corresponding to the preset emotional state; The fifth analysis subunit is used to perform weighted calculation on the concentration score corresponding to the behavior data, the concentration score corresponding to the reaction data, and the concentration score corresponding to the emotional state to obtain the concentration score corresponding to the behavior data.

7. The virtual simulation teaching management system based on big data according to claim 5 is characterized in that: The optimization unit comprises: The first optimization subunit is used to optimize the learning interaction process of students using a reinforcement learning algorithm according to the teaching strategy corresponding to each category of students and the students' attention fluctuation information, and obtain the students' state adjustment strategy; The second optimization subunit is used to analyze the switching frequency and task complexity of the trainee between different tasks according to the trainee's state adjustment strategy and task switching theory, and adjust the interactive task based on the switching frequency and task complexity between the different tasks to obtain an adjusted interactive task; The third optimization subunit is used to perform iterative genetic operations on the adjusted interactive tasks based on genetic algorithms, wherein the optimized interactive tasks corresponding to each student are obtained through iterative genetic operations, and the teaching strategies corresponding to each category of students and the optimized interactive tasks corresponding to each student are used as the optimized teaching strategies.

8. The virtual simulation teaching management system based on big data according to claim 5 is characterized in that: The processing unit comprises: The first processing sub-unit is used to perform state inference based on the student's concentration score and behavioral data using a Bayesian inference algorithm, wherein the student's current concentration state is predicted by updating the posterior probability distribution of the student's concentration; the second processing sub-unit is used to send the student's current concentration state to a preset learning intervention model, and formulate an intervention strategy for the rest of students in a concentration state, wherein the intervention strategy includes the student's rest time.

Images