Real-time teaching monitoring method and system based on cloud computing

Abstract

The invention discloses a real-time teaching monitoring method and system based on cloud computing, and relates to the technical field of education informationization, and the method comprises the steps: S101, obtaining real-time interaction data and learning behavior data generated in an online teaching process, the online teaching real-time interaction data comprises student answering response time, a video playing state and teacher-student voice interaction frequency, and the learning behavior data comprises a knowledge point mastering rate, learning duration distribution and an error question type. According to the invention, the intelligent real-time teaching monitoring platform is constructed, so that comprehensive monitoring, accurate analysis and dynamic adjustment of the teaching process are realized. The advanced cloud computing technology and algorithm optimization means are utilized, the system can collect and analyze teaching data in real time, teaching strategies and resource allocation are intelligently adjusted according to analysis results, and efficient development of teaching activities is ensured.

Description

Technical Field

[0001] This invention relates to the field of educational informatization technology, specifically to a real-time teaching monitoring method and system based on cloud computing. Background Technology

[0002] In the field of modern education, the widespread application of cloud computing technology provides crucial technical support for real-time teaching monitoring. By collecting classroom audio and video data and student behavior data through IoT devices, and combining this data with data from the school's teaching management platform for multi-dimensional analysis, existing technologies can already achieve basic assessment and dynamic monitoring of teaching quality.

[0003] However, in practical applications, there are still technical issues in the data transmission process that urgently need improvement. Particularly in large-scale educational settings such as universities, when multiple classrooms simultaneously upload large amounts of teaching data, including audio and video streams and behavioral data, campus networks are prone to congestion, especially during peak class times. Network congestion not only leads to data loss and transmission delays but can also result in incomplete or significantly reduced real-time teaching monitoring data, thus affecting the timeliness and accuracy of assessments of teaching progress. Furthermore, existing technologies often fail to adequately optimize for the characteristics of teaching data (such as data priority and time sensitivity), further exacerbating the irrationality of network resource allocation and low transmission efficiency. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a real-time teaching monitoring method and system based on cloud computing. It solves the problems of data delay and synchronization inconsistency that are easily caused by traditional polling or fixed-interval transmission mechanisms in high-concurrency student interaction scenarios, especially the problems of delayed teacher feedback and distorted teaching assessment caused by network fluctuations.

[0005] To achieve the above objectives, the present invention provides a real-time teaching monitoring method based on cloud computing, comprising: Step S101: Obtain real-time interactive data and learning behavior data generated during online teaching. Real-time interactive data includes student response time, video playback status, and teacher-student voice interaction frequency. Learning behavior data includes knowledge point mastery rate, learning time distribution, and types of incorrect questions. Step S102: Classify and process the real-time interactive data and learning behavior data of online teaching to obtain the data classification results. Based on the classification results, establish a dynamic priority queue and build a real-time teaching monitoring information processing mirror system in the cloud server. In the dynamic priority queue, preprocess and extract features from the real-time interactive data and learning behavior data to obtain interactive synchronicity features, learning focus features, and knowledge mastery progress features. Step S103: Retrieve historical teaching data, construct a first real-time teaching monitoring model using historical teaching data and interactive synchronicity features, learning focus features, and knowledge mastery progress features, and optimize the model in real time. Train the improved Long Short-Term Memory Network (LSTM) model using real-time interactive data and learning behavior data in the real-time teaching monitoring information processing mirror system to obtain a second real-time teaching monitoring model. Optimize the second real-time teaching monitoring model in real time in the real-time teaching monitoring information processing mirror system. Compare and analyze the prediction results of the first and second real-time teaching monitoring models. If the error between the two is less than a preset threshold, the first real-time teaching monitoring model takes effect. Step S104: Substitute real-time interactive data and learning behavior data into the first real-time teaching monitoring model to obtain the trend of student learning progress changes, collect real-time classroom information in the current teaching scenario, compare the real-time classroom information with the trend of student learning progress changes, and generate real-time teaching deviation value. Step S105: Analyze the real-time teaching deviation value, extract its corresponding timestamp information, establish a correlation between the timestamp information and the specific teaching link, and generate teaching content or strategy suggestions that need to be adjusted.

[0006] Preferably, step S101 includes: The system checks the completeness of real-time interactive data and learning behavior data. The check content includes student response time, video playback status, frequency of teacher-student voice interaction, knowledge point mastery rate, learning time distribution, and types of incorrect questions. The raw data check results are obtained, which include qualified data and missing data information. The raw data check results are compared with historical teaching data. If the data types match, the real-time interactive data and learning behavior data are matched with the historical teaching data. If the data types do not match, the data that differs between the raw data check results and the historical teaching data is retrieved and its data type is identified to obtain the information on the differing teaching data. New data detection categories are established based on differentiated teaching data. The data corresponding to the new data detection categories and the data corresponding to the original data detection results are substituted into the preset digital twin model to obtain a historical digital twin image of the teaching process. Obtain the coordinate information of teaching activities in the historical digital twin image of the teaching process. Use the known coordinates of the teaching base points to verify the accuracy of the coordinate information of teaching activities in the historical digital twin image of the teaching process. If the coordinate information of teaching activities in the historical digital twin image of the teaching process passes the verification, the real-time interactive data and learning behavior data will be effective and marked as usable data.

[0007] Preferably, step S102 includes: Verify whether real-time interactive data and learning behavior data have been classified according to predetermined classification criteria, which include data type, teaching subject and timestamp. Perform preprocessing on the classified data, including data cleaning, data standardization and data transformation. The system receives requests for teaching effectiveness evaluation. Based on these requests, it determines the types of features to be extracted and uses feature time series analysis to extract the selected features from the preprocessed data. The effectiveness of the extracted features is tested using historical data or known teaching effectiveness evaluation results. Based on the test results, the feature extraction strategy is adjusted, including modifying the feature type and optimizing the feature extraction algorithm.

[0008] Preferably, step S103 includes: The real-time interactive data and learning behavior data acquired in real time are simultaneously input into the first real-time teaching monitoring model and the second real-time teaching monitoring model; The first real-time teaching monitoring model and the second real-time teaching monitoring model make predictions based on the input data and generate corresponding prediction results, which include the first prediction result and the second prediction result. Compare the first prediction result and the second prediction result, and calculate the error between the first prediction result and the second prediction result; Define an error index to quantify the difference between the first and second prediction results, calculate the data processing result error of the first and second real-time teaching monitoring models, and compare the data processing result error with a preset error range. If the error of the real-time data processing result is within the preset error range, the first real-time teaching monitoring model is deemed to be effective; if the error exceeds the preset range, the first real-time teaching monitoring model and the second real-time teaching monitoring model are optimized, and the above steps are repeated for verification.

[0009] Preferably, step S104 includes: Before comparing real-time classroom information with the trend of student learning progress, a parallel data processing group is established. A data group correspondence is established between the real-time classroom information, the trend of student learning progress, and the parallel data processing group. Based on the data group correspondence, a data group processing task is established, and the data processing task is executed to perform parallel preprocessing of the real-time classroom information and the trend of student learning progress.

[0010] This invention also provides a cloud computing-based real-time teaching monitoring system, applying the aforementioned cloud computing-based real-time teaching monitoring method, comprising: a cloud server, teacher terminals, student terminals, and a management backend. The cloud server establishes communication connections with the teacher terminals, student terminals, and management backend respectively. The cloud server analyzes and processes data from the teacher terminals, student terminals, and management backend, generates teaching adjustment suggestions, and transmits them to each terminal. The cloud server includes: The data acquisition unit is used to acquire real-time interactive data and learning behavior data generated during online teaching. Real-time interactive data in online teaching includes student response time, video playback status, and frequency of teacher-student voice interaction. Learning behavior data includes knowledge point mastery rate, learning time distribution, and types of incorrect questions. The data classification unit is used to classify and process real-time interactive data and learning behavior data in online teaching, obtain data classification results, and establish a dynamic priority queue based on the classification results. In the dynamic priority queue, the real-time interactive data and learning behavior data are preprocessed and feature extracted to obtain interactive synchronicity features, learning focus features, and knowledge mastery progress features. The model building unit is used to retrieve historical teaching data and construct a first real-time teaching monitoring model by combining historical teaching data with interactive synchronicity features, learning focus features, and knowledge mastery progress features. This model is then optimized in real time. In the real-time teaching monitoring information processing mirror system, an improved Long Short-Term Memory (LSTM) network model is trained using real-time interactive data and learning behavior data to obtain a second real-time teaching monitoring model. This second real-time teaching monitoring model is then optimized in real time in the real-time teaching monitoring information processing mirror system. The prediction results of the first and second real-time teaching monitoring models are compared and analyzed. If the error between the two is less than a preset threshold, the first real-time teaching monitoring model becomes effective. The deviation analysis unit is used to input real-time interactive data and learning behavior data into the first real-time teaching monitoring model to obtain the trend of student learning progress, collect real-time classroom information in the current teaching scenario, compare the real-time classroom information with the trend of student learning progress, and generate real-time teaching deviation values. The strategy generation unit is used to analyze real-time teaching deviation values, extract their corresponding timestamp information, establish a correlation between the timestamp information and specific teaching links, and generate teaching content or strategy suggestions that need to be adjusted.

[0011] Preferably, the data acquisition unit is further used for: The training information data detection unit detects the completeness of real-time interactive data and learning behavior data. The detection content includes student response time, video playback status, frequency of teacher-student voice interaction, knowledge point mastery rate, learning time distribution, and error question types, obtaining raw data detection results. The raw data detection results include qualified data and missing data information. The raw data detection results are compared with historical teaching data in terms of data type. If the data type comparison is consistent, the real-time interactive data and learning behavior data are matched with the historical teaching data. If the data type comparison is inconsistent, the data that differs between the raw data detection results and the historical teaching data is retrieved and its data type is identified to obtain the difference teaching data information. The digital image simulation unit establishes new data detection categories based on differentiated teaching data information, and substitutes the data corresponding to the new data detection categories and the data corresponding to the original data detection results into the preset digital twin model to obtain historical digital twin images of the teaching process. The data logic detection unit acquires the coordinate information of teaching activities in the historical digital twin image of the teaching process. Using the known coordinates of the teaching base points, it verifies the accuracy of the coordinate information of teaching activities in the historical digital twin image of the teaching process. If the coordinate information of teaching activities in the historical digital twin image of the teaching process passes the verification, the real-time interactive data and learning behavior data become effective and are marked as usable data.

[0012] Preferably, the data classification unit is further used for: The data feature verification unit verifies whether real-time interactive data and learning behavior data have been classified according to predetermined classification criteria, which include data type, teaching subject, and timestamp. The unit also performs preprocessing on the classified data, including data cleaning, data standardization, and data transformation. The data feature extraction unit receives the requirements for teaching effectiveness evaluation, determines the types of features to be extracted based on the requirements, and uses the feature time series analysis method to extract the selected features from the preprocessed data. The data feature verification unit uses historical data or known teaching effectiveness evaluation results to verify the effectiveness of the extracted features. Based on the verification results, the feature extraction strategy is adjusted, including modifying the feature type and optimizing the feature extraction algorithm.

[0013] Preferably, the model building unit is further used for: The data output unit simultaneously inputs the real-time interactive data and learning behavior data acquired in real time into the first real-time teaching monitoring model and the second real-time teaching monitoring model. The data prediction unit, the first real-time teaching monitoring model and the second real-time teaching monitoring model respectively make predictions based on the input data and generate corresponding prediction results, including the first prediction result and the second prediction result; The real-time data comparison unit compares the first prediction result and the second prediction result, and calculates the error between the first prediction result and the second prediction result. The error definition unit defines an error index to quantify the difference between the first prediction result and the second prediction result, calculates the data processing result error of the first real-time teaching monitoring model and the second real-time teaching monitoring model, and compares the data processing result error with a preset error range. The model error judgment unit determines that the first real-time teaching monitoring model is effective if the error of the real-time data processing result is within the preset error range; if the error exceeds the preset range, it optimizes the first real-time teaching monitoring model and the second real-time teaching monitoring model, and repeats the above steps for verification.

[0014] Preferably, the strategy generation unit is further configured to: Before comparing real-time classroom information with the trend of student learning progress, the data parallel processing grouping unit establishes a data grouping correspondence between the real-time classroom information, the trend of student learning progress, and the data processing parallel processing grouping. Based on the data grouping correspondence, a data grouping processing task is established, and the data processing task is executed to perform parallel preprocessing of the real-time classroom information and the trend of student learning progress in groups. Beneficial effects

[0015] This invention optimizes data processing efficiency and reduces the impact of network congestion by utilizing dynamic priority queuing and cloud computing technologies, ensuring the integrity and real-time nature of monitoring data. A dual-model comparison and verification mechanism is constructed to improve the accuracy of teaching monitoring, precisely generating student learning trends and teaching deviation values. Teaching strategies are dynamically adjusted based on the analysis results, adapting to various modes such as online education and remote training, significantly improving teaching efficiency and quality. It solves the data latency and synchronization inconsistency problems of traditional transmission mechanisms in high-concurrency scenarios, avoiding delays in teacher feedback and distortion of teaching assessments caused by network fluctuations. Attached Figure Description

[0016] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] This invention provides a cloud-based real-time teaching monitoring method and system. Through cloud computing technology, it achieves comprehensive collection, classification, feature extraction, and dynamic modeling analysis of real-time interactive data and learning behavior data during online teaching, thereby generating precise teaching adjustment suggestions. The following is in conjunction with the appendix... Figure 1 The invention will also include detailed descriptions of the technical details mentioned in the invention description, and will provide a detailed explanation of the specific embodiments of the invention.

[0019] like Figure 1 As shown, the cloud-based real-time teaching monitoring system includes a cloud server, teacher terminals, student terminals, and a management backend. The cloud server establishes communication connections with the teacher terminals, student terminals, and management backend. The cloud server internally contains multiple functional modules, including a data acquisition unit, a data classification unit, a model building unit, a deviation analysis unit, and a strategy generation unit. These modules work collaboratively to complete the entire process from data acquisition to teaching optimization. In practical applications, the cloud server is deployed in the cloud using a distributed computing architecture, enabling it to efficiently handle high-concurrency data requests from multiple terminals, ensuring system stability and real-time performance.

[0020] First, the data acquisition unit is responsible for acquiring real-time interactive and learning behavior data generated during online teaching. Taking an online class as an example, assuming 50 students participate, each accessing the system through a student terminal, the data acquisition unit captures real-time interactive data such as students' response times, video playback status, and the frequency of teacher-student voice interactions. It also records learning behavior data such as knowledge point mastery rates, learning time distribution, and types of incorrect questions. All data is uploaded to the cloud server via teacher and student terminals and undergoes preliminary processing by the data acquisition unit. For example, at a certain moment, a student's response time is 3 seconds, the video playback status is "75% progress + not paused," and the voice interaction frequency is 2 times per minute. This information will be marked as the student's current learning status data and stored on the cloud server.

[0021] Next, the data classification unit classifies the collected data. Based on the nature of the data, real-time interactive data and learning behavior data are divided into different categories. For example, response time and voice interaction frequency belong to high-frequency updated real-time interactive data, while knowledge point mastery rate and learning time distribution belong to low-frequency updated learning behavior data. To improve subsequent processing efficiency, the data classification unit uses a dynamic priority queue to sort the data of different categories. Specifically, high-frequency updated real-time interactive data is given a higher priority to quickly pass it to subsequent processing stages, while low-frequency updated learning behavior data is arranged in a lower priority queue. Based on this, the data classification unit also performs preprocessing operations on the data, including: using the 3σ criterion to remove outliers in data such as response time and voice interaction frequency; using the K-nearest neighbor (K=5) algorithm to fill in missing knowledge point mastery rate data based on the average mastery rate of students in the same class; and applying Z-score standardization to all feature data. ,in The mean, Normalization was performed on the standard deviation.

[0022] Subsequently, the model building unit constructs a first real-time teaching monitoring model and a second real-time teaching monitoring model based on the preprocessed data. The construction process of the first real-time teaching monitoring model is as follows: First, historical teaching data is retrieved. This data is usually stored in the historical database of a cloud server, covering interaction data and learning behavior data from multiple previous teaching activities. Then, the model is trained using historical teaching data and currently extracted interaction synchronicity features, learning focus features, and knowledge mastery progress features. Interaction synchronicity features reflect the degree of matching between the frequency of student interaction in the classroom and the teacher's teaching pace; learning focus features measure the level of student attention concentration in the classroom; and knowledge mastery progress features describe the depth of students' understanding of knowledge points. Through the combination of these features, the first real-time teaching monitoring model can predict the trend of changes in students' learning status. At the same time, the model building unit also introduces an improved Long Short-Term Memory (LSTM) network model to train the real-time interaction data and learning behavior data to obtain the second real-time teaching monitoring model. The improved LSTM model achieves dynamic processing of time series data through forget gates, input gates, and output gates. The core formulas include: Forget Gate (controls the proportion of historical information retained): in, Here is the forget gate weight matrix. It is the Sigmoid activation function. , It is the concatenated vector of the previous hidden state and the current input. This is the forget gate bias term.

[0023] Input gate (controls the proportion of new information input): , in, , The input gate weight matrix, The input gate activation value, Candidate cell state, The hyperbolic tangent activation function is used. , This is the input gate bias term.

[0024] Cell status update: in, The current cell state, This represents the cell state at the previous moment.

[0025] Output gate (controls the current output information): in, This is the output gate weight matrix. The output gate activation value. This is the output gate bias term.

[0026] in, This represents the hidden state at the current moment, used to capture the temporal dependence of students' learning states; It indicates the hidden state at the previous moment, reflecting the influence of historical learning behavior; This represents the current input data, i.e., real-time interactive data or learning behavior data; and These are weight matrices for the hidden state and the input data, respectively, used to adjust the importance of different data. This is a bias term used to adjust the overall output range of the model; For activation functions, ReLU or Sigmoid functions are typically chosen to enhance nonlinear expressiveness. By weighting the combination of historical states and current inputs, the improved LSTM model can effectively capture the time dependence and dynamic changes in teaching data, thereby achieving accurate prediction of students' learning status.

[0027] After the model is built, the deviation analysis unit compares and analyzes the prediction results of the first and second real-time teaching monitoring models. Assume the first model predicts a student's knowledge mastery rate will be 80% in the next time period, while the second model predicts 75%. If the error between the two is less than a preset threshold (e.g., 5%), the first real-time teaching monitoring model is considered effective, and its prediction results will be used for subsequent teaching adjustments. Conversely, if the error exceeds the threshold, the model needs further optimization until the accuracy requirements are met.

[0028] After the model takes effect, the strategy generation unit inputs real-time interaction data and learning behavior data into the first real-time teaching monitoring model to obtain the trend of student learning progress. For example, suppose a student's knowledge mastery rate in the current class is 60%, and the model predicts that their mastery rate will increase to 70% in the next time period. At this time, the deviation analysis unit collects real-time classroom information in the current teaching scenario, such as the key content being explained by the teacher and the students' classroom participation, and compares this information with the trend of student learning progress. If it is found that the student's learning progress is lower than expected, a real-time teaching deviation value is generated. The formula for calculating the real-time teaching deviation value is: in The actual mastery rate of knowledge points (%) To predict the mastery rate (%) of knowledge points. This represents the real-time teaching deviation value (%).

[0029] This formula can quantify the gap between students' learning status and expected goals.

[0030] Finally, the strategy generation unit analyzes the real-time teaching deviation value, extracts its corresponding timestamp information, and establishes a correlation between the timestamp information and specific teaching segments. For example, suppose a student's real-time teaching deviation value is 10%, and the timestamp corresponding to this deviation value is the 30th minute of class. In this case, the strategy generation unit will analyze the teaching content of the 30th minute (corresponding to the "function differentiation" knowledge point), and combine it with the information in the learning behavior data that "the incorrect question type is 'composite function differentiation'," to determine that the student has gaps in their understanding of this knowledge point. Based on this, the system will generate corresponding teaching adjustment suggestions, such as prompting teachers to slow down the explanation pace, add case analysis, or provide additional practice questions. These suggestions are transmitted to the teacher's terminal via the cloud server for the teacher's reference and implementation.

[0031] Throughout the implementation process, the collaborative work between the cloud server, teacher terminals, student terminals, and the management backend is crucial. For example, when a teacher receives suggestions for adjusting teaching methods, they can adjust their teaching strategies accordingly and upload the revised content to the cloud server for students to view in real time. Furthermore, the management backend can access global teaching data through the cloud server to generate teaching evaluation reports, providing decision support for subsequent teaching activities.

[0032] In summary, this invention, by constructing an intelligent real-time teaching monitoring platform, achieves comprehensive monitoring, precise analysis, and dynamic adjustment of the teaching process. Utilizing advanced cloud computing technology and algorithm optimization methods, the system can collect and analyze teaching data in real time, intelligently adjusting teaching strategies and resource allocation based on the analysis results to ensure the efficient conduct of teaching activities. In practical applications, this invention can be widely used in various teaching models such as online education, remote training, and blended learning, significantly improving teaching efficiency and quality.

[0033] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0034] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A real-time teaching monitoring method based on cloud computing, characterized in that, include: Step S101: Obtain real-time interactive data and learning behavior data generated during online teaching. Real-time interactive data includes student response time, video playback status, and teacher-student voice interaction frequency. Learning behavior data includes knowledge point mastery rate, learning time distribution, and types of incorrect questions. Step S102: Classify and process the real-time interactive data and learning behavior data of online teaching to obtain the data classification results. Based on the classification results, establish a dynamic priority queue and build a real-time teaching monitoring information processing mirror system in the cloud server. In the dynamic priority queue, preprocess and extract features from the real-time interactive data and learning behavior data to obtain interactive synchronicity features, learning focus features, and knowledge mastery progress features. Step S103: Retrieve historical teaching data, construct a first real-time teaching monitoring model using historical teaching data and interactive synchronicity features, learning focus features, and knowledge mastery progress features, and optimize the model in real time. Train the improved Long Short-Term Memory Network (LSTM) model using real-time interactive data and learning behavior data in the real-time teaching monitoring information processing mirror system to obtain a second real-time teaching monitoring model. Optimize the second real-time teaching monitoring model in real time in the real-time teaching monitoring information processing mirror system. Compare and analyze the prediction results of the first and second real-time teaching monitoring models. If the error between the two is less than a preset threshold, the first real-time teaching monitoring model takes effect. Step S104: Substitute real-time interactive data and learning behavior data into the first real-time teaching monitoring model to obtain the trend of student learning progress changes, collect real-time classroom information in the current teaching scenario, compare the real-time classroom information with the trend of student learning progress changes, and generate real-time teaching deviation value. Step S105: Analyze the real-time teaching deviation value, extract its corresponding timestamp information, establish a correlation between the timestamp information and the specific teaching link, and generate teaching content or strategy suggestions that need to be adjusted.

2. The real-time teaching monitoring method based on cloud computing as described in claim 1, characterized in that, Step S101 includes: The system checks the completeness of real-time interactive data and learning behavior data. The check content includes student response time, video playback status, frequency of teacher-student voice interaction, knowledge point mastery rate, learning time distribution, and types of incorrect questions. The raw data check results are obtained, which include qualified data and missing data information. The raw data check results are compared with historical teaching data. If the data types match, the real-time interactive data and learning behavior data are matched with the historical teaching data. If the data types do not match, the data that differs between the raw data check results and the historical teaching data is retrieved and its data type is identified to obtain the information on the differing teaching data. New data detection categories are established based on differentiated teaching data. The data corresponding to the new data detection categories and the data corresponding to the original data detection results are substituted into the preset digital twin model to obtain a historical digital twin image of the teaching process. Obtain the coordinate information of teaching activities in the historical digital twin image of the teaching process. Use the known coordinates of the teaching base points to verify the accuracy of the coordinate information of teaching activities in the historical digital twin image of the teaching process. If the coordinate information of teaching activities in the historical digital twin image of the teaching process passes the verification, the real-time interactive data and learning behavior data will be effective and marked as usable data.

3. The real-time teaching monitoring method based on cloud computing as described in claim 1, characterized in that, Step S102 includes: Verify whether real-time interactive data and learning behavior data have been classified according to predetermined classification criteria, which include data type, teaching subject and timestamp. Perform preprocessing on the classified data, including data cleaning, data standardization and data transformation. The system receives requests for teaching effectiveness evaluation. Based on these requests, it determines the types of features to be extracted and uses feature time series analysis to extract the selected features from the preprocessed data. The effectiveness of the extracted features is tested using historical data or known teaching effectiveness evaluation results. Based on the test results, the feature extraction strategy is adjusted, including modifying the feature type and optimizing the feature extraction algorithm.

4. The real-time teaching monitoring method based on cloud computing as described in claim 1, characterized in that, Step S103 includes: The real-time interactive data and learning behavior data acquired in real time are simultaneously input into the first real-time teaching monitoring model and the second real-time teaching monitoring model; The first real-time teaching monitoring model and the second real-time teaching monitoring model make predictions based on the input data and generate corresponding prediction results, which include the first prediction result and the second prediction result. Compare the first prediction result and the second prediction result, and calculate the error between the first prediction result and the second prediction result; Define an error index to quantify the difference between the first and second prediction results, calculate the data processing result error of the first and second real-time teaching monitoring models, and compare the data processing result error with a preset error range. If the error of the real-time data processing result is within the preset error range, the first real-time teaching monitoring model is deemed to be effective; if the error exceeds the preset range, the first real-time teaching monitoring model and the second real-time teaching monitoring model are optimized, and the above steps are repeated for verification.

5. The real-time teaching monitoring method based on cloud computing as described in claim 1, characterized in that, Step S104 includes: Before comparing real-time classroom information with the trend of student learning progress, a parallel data processing group is established. A data group correspondence is established between the real-time classroom information, the trend of student learning progress, and the parallel data processing group. Based on the data group correspondence, a data group processing task is established, and the data processing task is executed to perform parallel preprocessing of the real-time classroom information and the trend of student learning progress.

6. A cloud-based real-time teaching monitoring system, employing the cloud-based real-time teaching monitoring method as described in any one of claims 1 to 5, characterized in that, include: The cloud server comprises a cloud server, teacher terminals, student terminals, and a management backend. The cloud server establishes communication connections with each of these terminals. It analyzes and processes data from these terminals, generates teaching adjustment suggestions, and transmits them to each terminal. The cloud server includes: The data acquisition unit is used to acquire real-time interactive data and learning behavior data generated during online teaching. Real-time interactive data in online teaching includes student response time, video playback status, and frequency of teacher-student voice interaction. Learning behavior data includes knowledge point mastery rate, learning time distribution, and types of incorrect questions. The data classification unit is used to classify and process real-time interactive data and learning behavior data in online teaching, obtain data classification results, and establish a dynamic priority queue based on the classification results. In the dynamic priority queue, the real-time interactive data and learning behavior data are preprocessed and feature extracted to obtain interactive synchronicity features, learning focus features, and knowledge mastery progress features. The model building unit is used to retrieve historical teaching data and construct a first real-time teaching monitoring model by combining historical teaching data with interactive synchronicity features, learning focus features, and knowledge mastery progress features. This model is then optimized in real time. In the real-time teaching monitoring information processing mirror system, an improved Long Short-Term Memory (LSTM) network model is trained using real-time interactive data and learning behavior data to obtain a second real-time teaching monitoring model. This second real-time teaching monitoring model is then optimized in real time in the real-time teaching monitoring information processing mirror system. The prediction results of the first and second real-time teaching monitoring models are compared and analyzed. If the error between the two is less than a preset threshold, the first real-time teaching monitoring model becomes effective. The deviation analysis unit is used to input real-time interactive data and learning behavior data into the first real-time teaching monitoring model to obtain the trend of student learning progress, collect real-time classroom information in the current teaching scenario, compare the real-time classroom information with the trend of student learning progress, and generate real-time teaching deviation values. The strategy generation unit is used to analyze real-time teaching deviation values, extract their corresponding timestamp information, establish a correlation between the timestamp information and specific teaching links, and generate teaching content or strategy suggestions that need to be adjusted.

7. The cloud-based real-time teaching monitoring system as described in claim 6, characterized in that, The data acquisition unit is also used for: The training information data detection unit detects the completeness of real-time interactive data and learning behavior data. The detection content includes student response time, video playback status, frequency of teacher-student voice interaction, knowledge point mastery rate, learning time distribution, and error question types, obtaining raw data detection results. The raw data detection results include qualified data and missing data information. The raw data detection results are compared with historical teaching data in terms of data type. If the data type comparison is consistent, the real-time interactive data and learning behavior data are matched with the historical teaching data. If the data type comparison is inconsistent, the data that differs between the raw data detection results and the historical teaching data is retrieved and its data type is identified to obtain the difference teaching data information. The digital image simulation unit establishes new data detection categories based on differentiated teaching data information, and substitutes the data corresponding to the new data detection categories and the data corresponding to the original data detection results into the preset digital twin model to obtain historical digital twin images of the teaching process. The data logic detection unit acquires the coordinate information of teaching activities in the historical digital twin image of the teaching process. Using the known coordinates of the teaching base points, it verifies the accuracy of the coordinate information of teaching activities in the historical digital twin image of the teaching process. If the coordinate information of teaching activities in the historical digital twin image of the teaching process passes the verification, the real-time interactive data and learning behavior data become effective and are marked as usable data.

8. The cloud computing-based real-time teaching monitoring system as described in claim 6, characterized in that, The data classification unit is also used for: The data feature verification unit verifies whether real-time interactive data and learning behavior data have been classified according to predetermined classification criteria, which include data type, teaching subject, and timestamp. The unit also performs preprocessing on the classified data, including data cleaning, data standardization, and data transformation. The data feature extraction unit receives the requirements for teaching effectiveness evaluation, determines the types of features to be extracted based on the requirements, and uses the feature time series analysis method to extract the selected features from the preprocessed data. The data feature verification unit uses historical data or known teaching effectiveness evaluation results to verify the effectiveness of the extracted features. Based on the verification results, the feature extraction strategy is adjusted, including modifying the feature type and optimizing the feature extraction algorithm.

9. The real-time teaching monitoring system based on cloud computing as described in claim 6, characterized in that, The model building unit is also used for: The data output unit simultaneously inputs the real-time interactive data and learning behavior data acquired in real time into the first real-time teaching monitoring model and the second real-time teaching monitoring model. The data prediction unit, the first real-time teaching monitoring model and the second real-time teaching monitoring model respectively make predictions based on the input data and generate corresponding prediction results, including the first prediction result and the second prediction result; The real-time data comparison unit compares the first prediction result and the second prediction result, and calculates the error between the first prediction result and the second prediction result. The error definition unit defines an error index to quantify the difference between the first prediction result and the second prediction result, calculates the data processing result error of the first real-time teaching monitoring model and the second real-time teaching monitoring model, and compares the data processing result error with a preset error range. The model error judgment unit determines that the first real-time teaching monitoring model is effective if the error of the real-time data processing result is within the preset error range; if the error exceeds the preset range, it optimizes the first real-time teaching monitoring model and the second real-time teaching monitoring model, and repeats the above steps for verification.

10. The real-time teaching monitoring system based on cloud computing as described in claim 6, characterized in that, The strategy generation unit is also used for: Before comparing real-time classroom information with the trend of student learning progress, the data parallel processing grouping unit establishes a data grouping correspondence between the real-time classroom information, the trend of student learning progress, and the data processing parallel processing grouping. Based on the data grouping correspondence, a data grouping processing task is established, and the data processing task is executed to perform parallel preprocessing of the real-time classroom information and the trend of student learning progress in groups.

Images