Intelligent teaching data management and optimization system

Abstract

The invention discloses an intelligent teaching data management and optimization system, and belongs to the technical field of education management systems. The system comprises a teaching data management platform, the teaching data management platform is in communication connection with a teaching multi-source data acquisition module, and the teaching multi-source data acquisition module is used for obtaining multi-source teaching data in a supervision area in real time; the system integrates teaching environment, student dynamic and resource state data, constructs a multi-dimensional analysis model, realizes second-level data acquisition, analysis and decision-making by means of edge calculation and stream data processing technologies, and intervenes the learning state of students in real time; the system applies a dynamic weight mechanism, adaptively adjusts a teaching strategy, intelligently optimizes resource allocation and can be self-repaired; in addition, the system implements differentiated teaching intervention based on the student risk level, guarantees the priority access of high-value resources through priority ranking, and remarkably improves the teaching pertinence and resource utilization efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of education management systems, and in particular to an intelligent teaching data management and optimization system. Background Art

[0002] The development of teaching management technology has always been closely linked to educational needs and technological progress. In the traditional teaching model, teachers rely on personal experience and manual observation to manage the classroom, lacking objective data support. At the end of the 20th century, with the popularization of computer technology, electronic tools such as student information management systems (SIS) and online learning platforms (LMS) began to be introduced into the education field, which initially realized the digital storage of teaching resources and basic data analysis. Entering the 21st century, breakthroughs in the Internet of Things, big data and artificial intelligence technologies have promoted the evolution of education management towards intelligence. The application of sensors, wearable devices and cloud computing has made it possible to collect teaching environment data, student behavior data and resource usage data in real time, providing a new technical foundation for teaching optimization.

[0003] Data on the teaching environment, student behavior, and resource status are scattered across different systems, lacking a unified integration mechanism. This leads to incomplete analysis results. For example, the lack of linkage between environmental data and student performance data makes it difficult to reveal the actual impact of temperature on attention. Most systems rely on offline batch analysis, which cannot capture transient anomalies in the classroom, such as sudden equipment failures or a sudden drop in student attention. Response delays can exacerbate teaching risks.

[0004] The above defects seriously restrict the practical application value of the intelligent teaching management system. There is an urgent need for an integrated solution that integrates multi-source data, supports dynamic analysis and can automatically optimize decision-making. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent teaching data management and optimization system to solve the problems raised.

[0006] To achieve the above objectives, the present invention provides the following technical solutions: an intelligent teaching data management and optimization system, the system comprising a teaching data management platform:

[0007] The teaching data management platform is communicatively connected to a teaching multi-source data acquisition module, which is used to acquire multi-source teaching data in the supervision area in real time. The multi-source teaching data includes classroom environment data packets, student dynamic data packets and learning resource data packets;

[0008] After receiving multi-source teaching data, the dynamic analysis model construction center builds a matching multi-dimensional analysis model based on the data packet type. Based on this, it obtains the analyzed classroom evaluation signal, student status warning signal and resource optimization requirements, and sends the analysis results to the management and control module and the intelligent optimization decision module;

[0009] After receiving classroom evaluation signals and resource optimization requirements, the intelligent optimization decision-making module generates a teaching strategy adjustment plan and a resource allocation priority list, and builds a data verification unit. It uses a cross-validation algorithm to perform confidence testing on the data used for the analysis results in the dynamic analysis model construction center, and determines whether to generate a data re-collection instruction based on the test results.

[0010] Furthermore, the dynamic analysis model building center analyzes the classroom environment data packet as follows:

[0011] The teaching multi-source data acquisition module continuously acquires classroom environment data packets based on the cycle from the initial to the current status of the school rules and curriculum progress, divides the cycle into sub-nodes according to the calendar, builds a timestamp archive, and retrieves the latest teaching environment data packets collected and stored in the recent timestamp archive. The teaching environment data packets include classroom room humidity, light intensity, and equipment operating status;

[0012] The equipment operating status includes the projector heat dissipation efficiency H proj and electronic whiteboard response delay T delay , projector heat dissipation efficiency H proj The linear ratio model for is:

[0013]

[0014] Where, ΔT proj It is expressed as the projector temperature rise value / ℃, which is collected in real time by the temperature sensor, t use Expressed as continuous usage time / hour, recorded by the device log. Model characteristics: Based on the direct linear relationship between physical quantities (temperature change and time), used to quantify the device's heat dissipation performance;

[0015] Electronic whiteboard response delay T delay The time difference calculation model is:

[0016] T delay =t touch -t display

[0017] Among them, t touch Expressed as touch input time, unit: milliseconds, t display Expressed as content display time, unit: milliseconds. Model characteristics: Directly calculate device response delay through timestamp difference, which belongs to basic timing analysis.

[0018] Furthermore, the dynamic analysis model construction center constructs the analysis process based on the student dynamic data packet as follows:

[0019] The student dynamic data packets are collected through the cameras and wearable devices in the classroom. The collected student dynamics are marked according to the time nodes, and a separate document storage and text "learning record" are constructed. The student posture features and eye activity probability P are extracted from the latest student dynamic data packets through CNN. eye , posture correct probability P posture , the attention concentration index is converted by the weighted fusion model:

[0020]

[0021] Among them, w i 、w j It is expressed as the weight coefficient, n is the number of sampling times, A focus It is expressed as the concentration index;

[0022] Get the number of questions N that the student answered correctly from the student's dynamic data package correct , and the total number of questions N that students participated in answering total , obtain the real-time answer accuracy rate through the online evaluation system:

[0023]

[0024] Among them, R correct Expressed as the real-time answer accuracy rate.

[0025] Furthermore, the workflow of the dynamic analysis model construction center for the classroom environment data package and the student dynamic data package is as follows:

[0026] The classroom environment data packets are normalized and the environmental quality index Q is generated through the multi-index normalization fusion model. env :

[0027]

[0028] Among them, T optimal Expressed as the optimal temperature, L max It is expressed as the maximum light intensity, α and β are expressed as environmental weights, and the pre-stored environmental threshold and environmental quality index Q are obtained from the teaching data management platform. env For comparison: If the environmental quality index Q env If the value is less than the environmental threshold, the environment is judged as not meeting the standard and a classroom evaluation signal is generated.

[0029] Furthermore, the learning state matrix M is constructed based on the student dynamic data packet. learn, retrieve the pre-stored attention concentration index lower threshold of 60%, answering accuracy threshold of 70%, and attention concentration index upper thresholds of 80% and 85% from the teaching data management platform, and input the support vector machine model through principal component analysis dimensionality reduction to obtain the student status warning signal. The student status warning signal is specifically:

[0030] If A focus <60% of the lower threshold of the concentration index, and R correct If the correct answer rate is less than 70%, it is judged to be high risk and marked as a high-risk warning level;

[0031] If the lower threshold of the concentration index is 60% ≤ A focus <Attention concentration index upper threshold 80%, and answering accuracy threshold 70% ≤ R correct If the concentration index is less than 85%, it is judged to be a medium risk and marked as a medium risk warning level;

[0032] The remaining situations are judged as low risk and are marked as low risk warning level.

[0033] Furthermore, the intelligent optimization decision module constructs a resource configuration priority list according to the resource status data packet as follows:

[0034] Calculate resource load rate based on resource status data packet: Among them L resource Expressed as resource load ratio;

[0035] Combined with classroom evaluation signals, define the priority coefficient: P priority =γ·Q env +δ·L resource , where P priority Expressed as the definition priority coefficient, γ and δ are expressed as Q env and L resource Dynamically adjust the weight, press P priority Generate a resource configuration list sorted from high to low.

[0036] Furthermore, the cross-validation method of the data validation unit includes:

[0037] The data set collected by the teaching multi-source data acquisition module is divided into a training set and a test set according to a ratio of 7:3, and is brought into the K-fold cross-validation calculation model obtained within the teaching data management platform to obtain the confidence C comf , retrieve the pre-stored intersection threshold and confidence C from the teaching data management platform comf Perform comparative analysis, if the confidence level C comf <intersection threshold, it is determined that there is an abnormality in data acquisition and a data re-acquisition instruction is generated; if the confidence level C comf>Intersection threshold, no signal is generated.

[0038] The beneficial effects of the present invention are:

[0039] 1. The present invention integrates three types of data, namely teaching environment, student dynamics and resource status, to construct a multi-dimensional analysis model. By correlating the room humidity and humidity of the classroom with the accuracy of students' answers, it accurately identifies the impact of environmental factors on learning outcomes, providing a comprehensive basis for formulating scientific teaching strategies. The multi-dimensional data is deeply integrated to break the information island. Relying on edge computing and stream data processing technology, the system can complete data collection, analysis and decision-making within seconds. When it detects that a student's attention index is continuously below the threshold, it pushes interactive questions to refocus their attention, avoiding the intervention lag caused by delays in traditional systems, and performing real-time dynamic analysis and immediate response.

[0040] 2. The present invention adopts a dynamic weight mechanism through the system, which can automatically adjust the strategy according to real-time data. It can give priority to ensuring the fluency of live courses when bandwidth is tight, and dynamically compress non-emergency resource loading, which not only meets core teaching needs but also improves resource utilization efficiency and adaptive intelligent decision-making optimization; the data verification unit continuously monitors the confidence of the model output, and immediately triggers the data re-collection instruction according to the deviation. When the environmental control strategy does not achieve the expected effect, the system automatically checks the sensor error or model parameter deviation, and completes self-repair, reducing the cost of manual intervention.

[0041] 3. The present invention implements differentiated intervention based on students' risk levels, providing high-risk students with micro-courses on knowledge points and intensive training on wrong questions, increasing the frequency of classroom questions and answers for medium-risk students, and opening up and expanding learning resources for low-risk students, thereby significantly improving the targeted teaching and precise teaching intervention; through dynamic priority sorting and intelligent allocation mechanism, it ensures priority access to high-value teaching resources and maximizes resource utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0043] Figure 1 It is a schematic diagram of the overall system flow of the present invention;

[0044] Figure 2 This is a schematic diagram of the connection structure between the teaching multi-source data acquisition module and the dynamic analysis model construction center of the present invention. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] Example 1: Please refer to Figure 1 - Figure 2 As shown, this embodiment is an intelligent teaching data management and optimization system, which includes a teaching data management platform. The teaching data management platform is communicatively connected to a teaching multi-source data acquisition module. The teaching multi-source data acquisition module is used to obtain multi-source teaching data in the supervision area in real time. The multi-source teaching data includes classroom environment data packets, student dynamic data packets and learning resource data packets.

[0047] After receiving multi-source teaching data, the dynamic analysis model construction center builds a matching multi-dimensional analysis model based on the data packet type. Based on this, it obtains the analyzed classroom evaluation signal, student status warning signal, and resource optimization requirements, and sends the analysis results to the management and control module and the intelligent optimization decision module. The dynamic analysis model construction center analyzes the classroom environment data packet as follows:

[0048] The teaching multi-source data acquisition module continuously obtains classroom environment data packets according to the cycle from the initial to the current status of the school regulations and curriculum progress, divides the cycle into sub-nodes according to the calendar, constructs a timestamp archive, and retrieves the latest teaching environment data packets collected and stored in the recent timestamp archive. The teaching environment data packets include classroom room and humidity, light intensity, and equipment operating status. It should be noted that the teaching environment data is collected through the temperature and humidity sensor / model DHT22, the light sensor / model BH1750, and the device log interface in JSON format, including timestamp, device ID, temperature and humidity values, and light intensity.

[0049] The equipment operating status includes the projector heat dissipation efficiency H proj and electronic whiteboard response delay T delay , projector heat dissipation efficiency H proj The linear ratio model for is:

[0050]

[0051] Where, ΔT proj Expressed as the projector temperature rise value / °C, collected in real time by the temperature sensor;

[0052] t use Expressed as continuous usage time / h, recorded by the device log;

[0053] Model characteristics of the linear ratio model: Based on the direct linear relationship between physical quantities, the physical quantity consists of temperature change and time, which is used to quantify the heat dissipation performance of the device;

[0054] Electronic whiteboard response delay T delay The time difference calculation model is:

[0055] T delay =t touch -t display

[0056] Among them, t touch Indicates the touch input time in milliseconds;

[0057] t display Indicates content display time, unit: milliseconds

[0058] Model characteristics of the time difference calculation model: Directly calculate the device response delay through the timestamp difference, which belongs to basic timing analysis.

[0059] The dynamic analysis model construction center constructs the analysis process based on the student dynamic data package as follows:

[0060] The student dynamic data packets are collected through the cameras and wearable devices in the classroom. The collected student dynamics are marked according to the time nodes, and a separate document storage and text "learning record" are constructed. The student posture features and eye activity probability P are extracted from the latest student dynamic data packets through CNN. eye , posture correct probability P posture It should be noted that the student dynamic data: the camera / resolution 1080p / 30fps captures the student's posture, the smart bracelet monitors the heart rate changes, the binary stream is parsed into a structured table by the edge computing node, among which CNN represents the convolutional neural network, and the attention concentration index is converted through the weighted fusion model:

[0061]

[0062] Among them, w i 、w j Expressed as the weight coefficient optimized by supervised learning, w i＝0.7 、w j＝0.3 , n represents the number of sampling times, 10 frames per second, A focus It is expressed as the concentration index;

[0063] Get the number of questions N that the student answered correctly from the student's dynamic data package correct , and the total number of questions N that students participated in answering total , obtain the real-time answer accuracy rate through the online evaluation system:

[0064]

[0065] Among them, R correct N represents the correct rate of students answering questions in class or online assessment, correct It is the number of questions answered correctly by the student, N total Represented as the total number of questions answered by students. This parameter is used to quantify students' mastery of knowledge points and is one of the core indicators for evaluating student status in the dynamic analysis model.

[0066] R correct Data Acquisition: Students' answering behavior is recorded in real time through online assessment systems, such as classroom interactive platforms and homework systems. Data including the correct answer to each question, the student's submitted answer, and the answer time stamp is obtained. The online assessment system is a pre-built and embedded subsystem within the teaching data management platform. It is pre-installed according to actual needs, but is not limited to this.

[0067] Data processing: The system automatically compares the answers and counts N correct and N total , calculate R according to the preset time window correct , such as single class, daily, and stored in student dynamic data packets.

[0068] The workflow of the dynamic analysis model construction center for classroom environment data packets and student dynamic data packets is as follows:

[0069] The classroom environment data packets are normalized and the environmental quality index Q is generated through the multi-index normalization fusion model. env :

[0070]

[0071] Among them, T optimal Expressed as the optimal temperature, L max It is expressed as the maximum light intensity, α and β are expressed as environmental weights, and the pre-stored environmental threshold and environmental quality index Q are obtained from the teaching data management platform. env For comparison, according to Q env Values automatically adjust device parameters:

[0072] If the environmental quality index Q env If the value is less than 70, the environmental threshold can be 70, and the full score is 100, then the environment is judged to be substandard, and a classroom evaluation signal is generated. The teacher evaluation signal is sent to the management and control module. After receiving the classroom evaluation signal, the management and control module triggers the control instruction. According to the parameters about the optimal temperature and maximum light intensity contained in the classroom evaluation signal, the pre-stored suitable threshold table in the teaching data management platform is retrieved, and its data is brought into the suitable threshold table. According to the data substituted into the data and the data in the suitable threshold table, the positive and negative differences are controlled, and the air conditioning target temperature is adjusted to Toptimal , T optimal The preset optimal temperature can be 25℃;

[0073] Curtain motor controls light to L max The maximum allowable light intensity is 500 lux;

[0074] Execution method: Sending instructions to smart terminals through IoT protocols, but not limited to this;

[0075] In the calculation of the environmental quality index, α = 0.6 and β = 0.4, focusing on the impact of temperature on teaching effectiveness. α and β are optimized through historical data regression analysis to ensure that the model adapts to the needs of different scenarios.

[0076] The learning state matrix M is constructed based on the student dynamic data packet learn , retrieve the pre-stored attention concentration index lower threshold of 60%, answer accuracy threshold of 70%, and attention concentration index upper thresholds of 80% and 85% from the teaching data management platform, and input them into the support vector machine / SVM model through principal component analysis / PCA dimensionality reduction to obtain the student status warning signal. The student status warning signal is specifically:

[0077] If A focus <60% of the lower threshold of the concentration index, and R correct If the correct answer rate is less than the threshold of 70%, it is determined to be high risk and marked as a high-risk warning level. After receiving the student status warning signal, which contains the high-risk warning level mark, the management and control module obtains the student data that generates the high-risk warning level and pushes a simplified version of the courseware and wrong question analysis to them;

[0078] If the lower threshold of the concentration index is 60% ≤ A focus <Attention concentration index upper threshold 80%, and answering accuracy threshold 70% ≤ R correct If the concentration index is less than 85%, it is determined to be medium risk and marked as a medium risk warning level. After receiving the student status warning signal containing the medium risk warning level mark, the management and control module will obtain the students with the data generating the medium risk warning level and increase the frequency of classroom interaction for them.

[0079] The rest of the cases are judged as low risk and marked as low risk warning level. After receiving the student status warning signal, which contains the low risk warning level mark, the management and control module obtains students with data that generates medium risk warning level and pushes them to provide extended learning materials.

[0080] Implementation method: Send content recommendations to student terminals through the education platform API;

[0081] The student status classification threshold is set based on educational psychology research to ensure the accuracy of early warning. After receiving the optimized content, the student terminal will feedback the learning time and performance improvement data to the system, execute the teaching strategy adjustment plan, and adjust the teaching equipment parameters and resource allocation strategy in real time. The system realizes closed-loop optimization of the teaching process through multi-source data fusion and dynamic feedback mechanism.

[0082] Example 2: This example is an intelligent teaching data management and optimization system. After receiving classroom evaluation signals and resource optimization requirements, the intelligent optimization decision-making module generates a teaching strategy adjustment plan and a resource allocation priority list, and constructs a data verification unit. It uses a cross-validation algorithm to perform confidence testing on the data used for the analysis results in the dynamic analysis model construction center, and determines whether to generate a data re-collection instruction based on the test results.

[0083] Resource status data: The server API interface obtains textbook click-through rate and network traffic data in real time, stores key-value pairs, and classifies them by resource type. The intelligent optimization decision module constructs a resource configuration priority list based on the resource status data packet as follows:

[0084] Calculate resource load rate based on resource status data packet:

[0085]

[0086] Among them L resource Expressed as resource load ratio;

[0087] Combined with classroom evaluation signals, the priority coefficient is defined by a linear weighted decision model:

[0088] P priority =γ·Q env +δ·L resource

[0089] Among them, P priority Expressed as the definition priority coefficient, γ and δ are expressed as Q env and L resource Dynamically adjust the weights, γ and δ are optimized through historical data regression analysis to ensure that the model adapts to the needs of different scenarios. priority Generate a resource allocation list from high to low. It should be noted that the model characteristics of the linear weighted decision model are: the environmental quality index Q env and resource load rate L resource According to the dynamic weights γ and δ, the resource allocation priority is generated, γ = 0.5, δ = 0.5, C total Indicated as the currently occupied bandwidth 50Mbps, such as the currently occupied bandwidth, data source: obtained in real time through the server monitoring interface, network traffic sensor or storage system log; C totalIf the total bandwidth is 100Mbps, then L resource＝50％ It should be noted that C total It is the core parameter in resource load rate calculation, indicating the real-time usage of a resource. Its specific meaning depends on the resource type.

[0090] When P priority When the value is greater than 90, bandwidth is allocated to live course resources first. 90 is the pre-stored allocation threshold limit retrieved from the teaching data management platform and sorted by priority coefficient. The specific operation is as follows:

[0091] Press P priority List dynamically allocated resources:

[0092] Bandwidth is allocated first to high-priority resources, such as live courses;

[0093] Limit or delay the loading of low-priority resources, such as document downloads;

[0094] Implementation method: Adjust routing policies through the software-defined network (SDN) controller.

[0095] The cross-validation method of the data validation unit includes:

[0096] The data set collected by the teaching multi-source data acquisition module is divided into a training set and a test set in a ratio of 7:3. 70% of the training set is used for model training and optimization of parameters such as the kernel function weight of SVM. The 30% of the test set is used to verify the generalization ability of the model and simulate the prediction effect in real scenarios. It is divided in chronological order, such as the first 7 days of data training and the last 3 days of data testing, to avoid future data leakage into the training process. It is brought into the K-fold cross-validation calculation model obtained within the teaching data management platform to obtain the confidence C comf , K-fold cross validation calculation model features, K = 5: based on the mean absolute error MAE between the predicted value and the true value, the confidence level is calculated to evaluate the reliability of the model. The pre-stored cross threshold and confidence level C are retrieved from the teaching data management platform. comf For comparison analysis, the intersection threshold can be set to 0.85:

[0097] If the confidence level C comf If the value is less than 0.85, the data acquisition is judged to be abnormal, indicating that the model has unreliable factors or data anomalies. A data re-collection instruction is generated, triggering data re-collection and model retraining. The model is retrained using historical data to train the SVM model. The kernel function is radial basis function / RBF. The hyperparameters are optimized through grid search. PCA dimensionality reduction retains 85% of the variance. The principal components are extracted as model inputs. The backup sensor is called to complete the data. The model is retrained and the parameters are updated.

[0098] Execution method: An automated script triggers the data pipeline and training tasks, performs K-fold cross-validation every 24 hours, updates model parameters, and automatically calls for backup sensor data when the confidence level falls below a threshold.

[0099] If the confidence level C comf >0.85, no signal is generated.

[0100] In combination with Example 1 and Example 2, the present invention focuses on smart teaching and creates an innovative teaching system. The system integrates three types of data: teaching environment, student dynamics, and resource status, builds a multi-dimensional analysis model, and uses edge computing and stream data processing technology to achieve second-level data collection, analysis and decision-making, and instantly intervene in students' learning status.

[0101] At the same time, the system uses a dynamic weight mechanism to adaptively adjust teaching strategies, intelligently optimize resource allocation, and can self-repair; in addition, the system implements differentiated teaching interventions based on students' risk levels, and ensures priority access to high-value resources through priority sorting, significantly improving the targeted nature of teaching and resource utilization efficiency.

[0102] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the scope of protection of the present invention.

[0103] In the description of this specification, the descriptions with reference to the terms "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. Related accessories include couplings, screws, gears, gaskets and other commonly used mechanical connection components in this field, but are not limited to these. They are replaced and adapted according to actual use.

[0104] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent teaching data management and optimization system, characterized in that: The system includes a teaching data management platform: The teaching data management platform is communicatively connected to a teaching multi-source data acquisition module, which is used to acquire multi-source teaching data in the supervision area in real time. The multi-source teaching data includes classroom environment data packets, student dynamic data packets and learning resource data packets; After receiving multi-source teaching data, the dynamic analysis model construction center builds a matching multi-dimensional analysis model based on the data packet type. Based on this, it obtains the analyzed classroom evaluation signal, student status warning signal and resource optimization requirements, and sends the analysis results to the management and control module and the intelligent optimization decision module; After receiving classroom evaluation signals and resource optimization requirements, the intelligent optimization decision-making module generates a teaching strategy adjustment plan and a resource allocation priority list, and builds a data verification unit. It uses a cross-validation algorithm to perform confidence testing on the data used for the analysis results in the dynamic analysis model construction center, and determines whether to generate a data re-collection instruction based on the test results.

2. The intelligent teaching data management and optimization system according to claim 1, characterized in that: The dynamic analysis model construction center analyzes the classroom environment data packet as follows: The teaching multi-source data acquisition module continuously acquires classroom environment data packets based on the cycle from the initial to the current status of the school rules and curriculum progress, divides the cycle into sub-nodes according to the calendar, builds a timestamp archive, and retrieves the latest teaching environment data packets collected and stored in the recent timestamp archive. The teaching environment data packets include classroom room humidity, light intensity, and equipment operating status; The equipment operating status includes the projector heat dissipation efficiency H proj and electronic whiteboard response delay T delay , projector heat dissipation efficiency H proj The linear ratio model for is: Where ΔT proj Expressed as the projector temperature rise value, t use Expressed as continuous usage time; Electronic whiteboard response delay T delay The time difference calculation model is: T delay =t touch -t display Among them, t touch Expressed as touch input time, t display Indicates the content display time.

3. The intelligent teaching data management and optimization system according to claim 2, characterized in that: The dynamic analysis model construction center constructs the analysis process based on the student dynamic data package as follows: Student dynamic data packets are collected through cameras and wearable devices in the classroom. The collected student dynamics are marked according to time nodes, and a separate document storage and text "learning activity record" label are constructed. From the latest student dynamic data packet, CNN is used to extract student posture features and eye activity probability P eye , posture correct probability P posture , the attention concentration index is converted by the weighted fusion model: Among them, w i 、w j It is expressed as the weight coefficient, n is the number of sampling times, A focus It is expressed as the concentration index; Get the number of questions N that the student answered correctly from the student's dynamic data package correct , and the total number of questions N that students participated in answering total , obtain the real-time answer accuracy rate through the online evaluation system: Among them, R correct Expressed as the real-time answer accuracy rate.

4. The intelligent teaching data management and optimization system according to claim 3, characterized in that: The workflow of the dynamic analysis model construction center for classroom environment data packets and student dynamic data packets is as follows: The classroom environment data packets are normalized and the environmental quality index Q is generated through the multi-index normalization fusion model. env : Among them, T optimal Expressed as the optimal temperature, L max It is expressed as the maximum light intensity, α and β are expressed as environmental weights, and the pre-stored environmental threshold and environmental quality index Q are obtained from the teaching data management platform. env For comparison: If the environmental quality index Q env If the value is less than the environmental threshold, the environment is judged as not meeting the standard and a classroom evaluation signal is generated.

5. The intelligent teaching data management and optimization system according to claim 4, characterized in that: The learning state matrix M is constructed based on the student dynamic data packet learn , retrieve the pre-stored attention concentration index lower threshold of 60%, answering accuracy threshold of 70%, and attention concentration index upper thresholds of 80% and 85% from the teaching data management platform, and input the support vector machine model through principal component analysis dimensionality reduction to obtain the student status warning signal. The student status warning signal is specifically: If A focus <60% of the lower threshold of the concentration index, and R correct If the correct answer rate is less than 70%, it is judged to be high risk and marked as a high-risk warning level; If the lower threshold of the concentration index is 60% ≤ A focus <Attention concentration index upper threshold 80%, and answering accuracy threshold 70% ≤ R correct If the concentration index is less than 85%, it is judged to be a medium risk and marked as a medium risk warning level; The remaining situations are judged as low risk and are marked as low risk warning level.

6. The intelligent teaching data management and optimization system according to claim 1, characterized in that: The intelligent optimization decision module constructs the resource configuration priority list according to the resource status data packet as follows: Calculate resource load rate based on resource status data packet: Among them L resource Expressed as resource load ratio; Combined with classroom evaluation signals, define the priority coefficient: P priority =γ·Q env +δ·L resource , where P priority Expressed as the definition priority coefficient, γ and δ are expressed as Q env and L resource Dynamically adjust the weight, press P priority Generate a resource configuration list sorted from high to low.

7. The intelligent teaching data management and optimization system according to claim 1, characterized in that: The cross-validation method of the data validation unit includes: The data set collected by the teaching multi-source data acquisition module is divided into a training set and a test set according to a ratio of 7:3, and is brought into the K-fold cross-validation calculation model obtained within the teaching data management platform to obtain the confidence C conf , and then retrieve the pre-stored intersection threshold and confidence C from the teaching data management platform comf Perform comparative analysis, if the confidence level C comf <intersection threshold, it is determined that there is an abnormality in data acquisition and a data re-acquisition instruction is generated; if the confidence level C comf >Intersection threshold, no signal is generated.

Images