Real-time dynamic learning data monitoring and visualization platform and method

Abstract

The invention relates to the technical field of teaching visualization, in particular to a real-time dynamic learning data monitoring and visualization platform and method. The platform comprises a teaching state docking acquisition reporting module, a teaching state modeling analysis module, a teaching recommendation before and after analysis module and a teaching supervision evaluation visualization module. A corresponding teaching basic state acquisition model can be constructed by combining teaching affair visual data docking, learning examination data docking and teaching environment data docking with a teaching basic state data warehouse, so that corresponding dynamic learning data can be synchronously monitored and acquired in real time and reported; teaching modeling feature analysis and personalized teaching analysis are carried out on dynamic learning data in a teaching basic state data warehouse, and meanwhile, student knowledge advancing and retreating analysis and teaching supervision evaluation visualization are carried out, so that a corresponding classroom teaching supervision visualization result is generated. According to the invention, the teaching quality can be improved, and the teaching supervision evaluation data can be presented in a visual and visualized manner.

Description

Technical Field

[0001] The present invention relates to the technical field of teaching visualization, and in particular to a real-time dynamic learning data monitoring and visualization platform and method. Background Art

[0002] With the development of big data and artificial intelligence technology, the field of education has gradually begun to explore new models of teaching supervision based on data analysis. By collecting, analyzing and visualizing multi-dimensional data in the teaching process, it can help schools and education management departments to achieve comprehensive monitoring of teachers' teaching behaviors, students' learning conditions, and the use of teaching resources. However, existing teaching supervision and evaluation methods usually rely on manual operations and regular data analysis. For example, traditional classroom observations and teaching feedback mainly rely on the subjective judgment of teachers and supervisors. The evaluation results are easily affected by personal experience, emotions and other factors, and the evaluation time point is relatively fixed, failing to fully track the dynamic changes in the teaching process. In addition, although the evaluation based on student performance can reflect part of the teaching effect, it does not fully cover all aspects of the teaching process, and lacks a timely data monitoring, reporting and feedback mechanism. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide a real-time dynamic learning data monitoring and visualization method and platform to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a real-time dynamic learning data monitoring and visualization platform is proposed, which includes the following modules: The teaching status docking collection and reporting module is used to build a corresponding teaching basic status collection model through the combination of teaching affairs visual data docking, learning and examination affairs data docking, and teaching environment data docking with the teaching basic status data warehouse; the corresponding dynamic learning data is collected and monitored in real time and synchronously by using the docking corresponding to the teaching basic status collection model, and reported to the corresponding teaching basic status data warehouse; The teaching state modeling and analysis module is used to perform teaching modeling feature analysis on dynamic learning data in the teaching basic state data warehouse to obtain a classroom teaching visual behavior feature set, a student learning behavior feature set, and a student environmental influencing factor set; based on the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environmental influencing factor set, a personalized teaching analysis is performed on the corresponding students to generate personalized teaching suggestions for classroom students; The pre- and post-teaching recommendation analysis module is used to obtain the students' learning performance and class participation before and after the teaching recommendation based on the personalized teaching suggestions for the students in the classroom, and to analyze the students' knowledge progress and regression on their learning performance before and after the teaching recommendation to obtain the students' knowledge learning progress and regression rate before and after the teaching recommendation; The teaching supervision and evaluation visualization module is used to visualize the teaching supervision and evaluation of students' knowledge learning progress and regression rates and classroom participation before and after teaching recommendations based on dynamic learning data, so as to generate corresponding classroom teaching supervision visualization results.

[0005] Furthermore, the teaching status docking collection and reporting module includes the following functions: By connecting the teaching affairs visual data, learning and examination affairs data, and teaching environment data with the teaching basic status data warehouse, a corresponding teaching basic status acquisition model is constructed; By configuring the corresponding teaching status synchronization collection time point, and based on the teaching status synchronization collection time point, using the corresponding docking of the teaching basic status collection model to monitor and collect the corresponding dynamic learning data in real time, including classroom dynamic visual data, learning dynamic behavior data and teaching dynamic environment data; Stream-process the corresponding dynamic learning data to generate a basic teaching state data stream; The basic teaching status data stream is transmitted and reported to the data sub-database in the basic teaching status data warehouse corresponding to the basic teaching status acquisition model.

[0006] Furthermore, the real-time synchronous monitoring and acquisition of the corresponding dynamic learning data based on the teaching state synchronous acquisition time point using the teaching basic state acquisition model corresponding to the docking includes: Based on the teaching status synchronous collection time point, the teaching affairs visual data corresponding to the basic teaching status collection model is used to connect the real-time monitoring collection of the corresponding classroom dynamic visual data, including the classroom panorama, teacher teaching activities and student classroom performance corresponding video stream data; Based on the teaching status synchronous collection time point, the learning and examination data corresponding to the teaching basic status collection model is connected to the real-time monitoring and collection of the corresponding learning dynamic behavior data, including the number of course visits, learning time, homework submission time, online test scores, students' corresponding body movements and sitting status in class; Based on the teaching status synchronous collection time point, the teaching environment data corresponding to the teaching basic status collection model is used to connect the real-time monitoring and collection of the corresponding teaching dynamic environment data, including the teacher's teaching content, students' questions and answers, classroom discussion sound corresponding voice text data and classroom learning environment data. The classroom learning environment data includes the corresponding temperature, humidity and light intensity of the classroom.

[0007] Furthermore, the teaching status modeling and analysis module includes the following functions: By combining convolutional neural networks and recurrent neural networks in the teaching basic status data warehouse to build a corresponding classroom visual behavior analysis model, and obtaining the corresponding classroom teacher and student behavior annotation data, the classroom visual behavior analysis model is trained based on the classroom teacher and student behavior annotation data to generate a classroom behavior analysis model that can identify the corresponding teacher and student behaviors; Input the classroom dynamic visual data into the classroom behavior analysis model that can identify the corresponding teacher and student behaviors to perform classroom visual behavior feature analysis, so as to identify the corresponding teaching behavior features of teachers such as explanation, demonstration and questioning, and the corresponding classroom behavior features of students such as attentive listening, whispering, discussion and absent-mindedness, and obtain the classroom teaching visual behavior feature set; The spatiotemporal convolutional network is used to analyze the spatiotemporal variation characteristics of the course access times, learning time, homework submission time, and online test scores corresponding to the learning dynamic behavior data, so as to analyze the variation characteristics of the corresponding student behaviors in the time and space sequence, and the student concentration statistics are performed on the corresponding body movements and sitting postures of the students in the classroom in the learning dynamic behavior data to obtain the corresponding concentration of the students in the classroom. At the same time, the variation characteristics of the student behaviors in the time and space sequence and the corresponding concentration of the students in the classroom are combined to obtain the student learning behavior feature set; Based on the dynamic teaching environment data, the learning impact assessment of students' corresponding concentration in the classroom is carried out to determine the environmental factors that affect students' concentration during the teacher's classroom teaching process, so as to obtain the student environmental influence factor set; key mining and dimensionality reduction processing are carried out between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influence factor set, so as to generate the dimensionality reduction feature set of the basic teaching state; Through the dimensionality reduction feature set analysis of the basic teaching status, the corresponding classroom learning evaluation of the students is obtained, including the corresponding learning history, interest preferences and knowledge mastery of the students. Personalized teaching analysis is performed based on the corresponding classroom learning evaluation of the students, so as to recommend corresponding learning resources and learning paths for the students' weak knowledge points and generate personalized teaching suggestions for students in the classroom.

[0008] Furthermore, the student concentration statistics of the corresponding body movements and sitting postures of the students in the classroom in the learning dynamic behavior data include: Performing time-series synchronization processing on the corresponding body movements and sitting postures of students in the classroom in the learning dynamic behavior data to generate corresponding student body movement sequences and student sitting posture sequences within the same time-series range; The corresponding student head rotation angle and student arm swing angle are obtained through the student body movement sequence, and the center of gravity shift analysis is performed according to the student head rotation angle and the student arm swing angle to obtain the degree of student center of gravity shift; The duration of the students' gaze at the blackboard is obtained through the student's body movement sequence, and the number of twists of the sitting body is obtained through the student's sitting state sequence; The sitting posture stability evaluation is calculated based on the duration of the students staring at the blackboard and the number of twists of the body in the sitting position, so as to obtain the degree of students' sitting posture stability; The students' concentration is counted based on the degree of their center of gravity shift and the stability of their sitting posture to obtain the corresponding concentration of the students in class.

[0009] Furthermore, the learning impact assessment and judgment of the students' corresponding concentration in class based on the teaching dynamic environment data includes: Based on the various environmental factors in the teaching dynamic environment data, a linear correlation impact assessment is performed on the corresponding concentration of students in the classroom to obtain the linear correlation influence coefficient between each teaching environment factor and concentration; Based on the linear correlation influence coefficient between each teaching environment factor and the concentration, the learning impact judgment and screening of each environmental factor in the teaching dynamic environment data is carried out. If the linear correlation influence coefficient between the corresponding teaching environment factor and the concentration is positively correlated, then the corresponding teaching environment factor is the environmental factor that affects the student's concentration; if the linear correlation influence coefficient between the corresponding teaching environment factor and the concentration is negatively correlated, then the corresponding teaching environment factor is not the environmental factor that affects the student's concentration. In this way, the environmental factors that affect the student's concentration during the teacher's classroom teaching process are judged and screened to obtain the student environment influencing factor set.

[0010] Furthermore, the key mining and dimensionality reduction processing between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influencing factor set includes: The correlation measurement calculation is performed between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influencing factor set, so as to quantitatively calculate the correlation coefficient between the two sub-features and form a corresponding correlation matrix to generate the basic state feature correlation matrix of teaching; Based on the correlation matrix of basic teaching status characteristics, the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influencing factor set are screened for behavioral feature key mining. If the correlation coefficient between the two sub-features in the correlation matrix of basic teaching status characteristics is greater than 0.75, it is considered that the two sub-features are all key features; if the correlation coefficient between the two sub-features in the correlation matrix of basic teaching status characteristics is equal to 0.75, it is considered that there is a key feature in the two sub-features; if the correlation coefficient between the two sub-features in the correlation matrix of basic teaching status characteristics is less than 0.75, it is considered that the two sub-features are not key features; select the corresponding key features according to the above judgment rules to generate the key feature set of basic teaching status; The principal component dimensionality reduction processing is performed on the corresponding sub-features in the key feature set of the basic teaching state to generate a dimensionality reduction feature set of the basic teaching state.

[0011] Furthermore, the teaching recommendation before and after analysis module includes the following functions: Based on the personalized teaching suggestions for students in class, the corresponding classroom teaching recommendations are implemented, and the corresponding learning performance and classroom participation of students before and after the teaching recommendations are obtained, where the learning performance includes the learning performance corresponding to each subject knowledge; Perform learning progress and regression space analysis on each subject knowledge within the students' corresponding learning scores before and after the teaching recommendation, so as to quantify the difference in learning scores of each subject knowledge before and after the teaching recommendation, and obtain the learning progress and regression space corresponding to each subject knowledge; The learning progress and regression range of each subject knowledge in the students' corresponding learning scores before and after the teaching recommendation is calculated, so as to quantify the ratio of the corresponding learning scores of each subject knowledge before and after the teaching recommendation, and obtain the learning progress and regression range corresponding to each subject knowledge; The students' knowledge progress and regression are summarized and calculated based on the learning progress and regression space and the learning progress and regression amplitude corresponding to each subject knowledge, so as to obtain the students' corresponding knowledge learning progress and regression rate before and after the teaching recommendation.

[0012] Furthermore, the teaching supervision and evaluation visualization module includes the following functions: According to the classroom dynamic visual data, learning dynamic behavior data and teaching dynamic environment data corresponding to the dynamic learning data, a teaching supervision evaluation index system is constructed to generate corresponding teacher teaching quality evaluation indicators, student learning effect evaluation indicators and teaching environment evaluation indicators. The teacher teaching quality evaluation indicators include the effectiveness of teaching methods, classroom interaction effects and the improvement of student learning outcomes. The student learning effect evaluation indicators include academic performance and learning interest cultivation. The teaching environment evaluation indicators include classroom learning impact, classroom facility applicability and environmental learning atmosphere. Based on the scores of the teacher teaching quality evaluation indicators, student learning effect evaluation indicators and teaching environment evaluation indicators, a teaching supervision data report is established for the students' knowledge learning progress and regression rate and classroom participation before and after the teaching recommendation, so as to generate a teacher teaching supervision quality data report; Perform linear analysis and visualization on the relationship between each indicator score and the knowledge learning progress or classroom participation rate in the teacher teaching supervision quality data report, so as to establish a linear visualization chart between each indicator score and the knowledge learning progress or classroom participation rate, so as to generate the corresponding classroom teaching supervision visualization results.

[0013] Furthermore, the present invention also provides a real-time dynamic learning data monitoring and visualization method, which is implemented based on the real-time dynamic learning data monitoring and visualization platform as described above, and the real-time dynamic learning data monitoring and visualization method includes: Through the combination of teaching visual data docking, learning and examination data docking, and teaching environment data docking and teaching basic status data warehouse, a corresponding teaching basic status collection model is constructed; the corresponding dynamic learning data is collected and monitored in real time and synchronously by using the docking corresponding to the teaching basic status collection model, and reported to the corresponding teaching basic status data warehouse; By conducting teaching modeling feature analysis on dynamic learning data in the teaching basic status data warehouse, a classroom teaching visual behavior feature set, a student learning behavior feature set, and a student environmental influencing factor set are obtained; based on the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environmental influencing factor set, personalized teaching analysis is conducted on the corresponding students to generate personalized teaching suggestions for classroom students; Based on the personalized teaching suggestions for students in class, the corresponding learning performance and classroom participation of students before and after the teaching recommendation are obtained, and the students' knowledge progress and regression analysis is performed on the corresponding learning performance before and after the teaching recommendation to obtain the corresponding knowledge learning progress and regression rate of students before and after the teaching recommendation; Based on dynamic learning data, the teaching supervision evaluation is visualized for students' knowledge learning progress and regression rates and classroom participation before and after teaching recommendations to generate corresponding classroom teaching supervision visualization results.

[0014] Beneficial effects of the present invention: The real-time dynamic learning data monitoring and visualization platform proposed in the present invention is generally composed of a teaching status docking collection and reporting module, a teaching status modeling and analysis module, a teaching recommendation before and after analysis module, and a teaching supervision and evaluation visualization module. Compared with the prior art, the beneficial effect of the present application is that through the organic combination of teaching affairs visual data docking, learning examination affairs data docking, and teaching environment data docking, a basic teaching status collection model can be accurately constructed. The core goal of this model is to comprehensively monitor and record various dynamic factors of classroom teaching, such as students' learning behavior, changes in the teaching environment, and the use of teaching resources. By docking, real-time synchronous monitoring and collection of these dynamic learning data, and reporting them to the basic teaching status data warehouse, high-quality original data support can be provided for subsequent data analysis. The greatest advantage of this process is the real-time and comprehensiveness of data collection, so that all teaching activities can be reflected in the data warehouse in real time, improving the accuracy and quality of teaching activities, so that teachers and managers can obtain a more comprehensive and true classroom status, avoiding the errors of traditional manual records in the past, thereby enabling timely data monitoring and reporting. Secondly, by conducting teaching modeling feature analysis on dynamic learning data in the teaching basic status data warehouse, we can extract key classroom teaching behavior characteristics, student learning behavior characteristics and the impact of environmental factors on learning. Based on these multi-dimensional characteristics, we can accurately locate students' personalized teaching needs, formulate specific teaching plans and personalized suggestions, maximize the learning needs of different students, and improve learning outcomes. This data-driven personalized teaching analysis can make the teaching process more flexible and efficient, and enhance students' learning enthusiasm and sense of participation, so that teachers can better adjust teaching strategies and promote students' independent learning and ability improvement. Then, by obtaining the changes in students' academic performance and classroom participation before and after the teaching recommendation based on personalized teaching suggestions for students in the classroom, a comprehensive evaluation of the teaching effect can be made. Specifically, changes in students' academic performance can reflect whether the teaching recommendation has effectively improved students' academic ability, while classroom participation can reflect the students' degree of involvement in the teaching content. By analyzing the progress and regression of students' academic performance before and after the teaching recommendation, the knowledge learning progress of each student after the teaching adjustment can be accurately evaluated. This analysis not only reveals the changes in students' knowledge mastery, but also helps teachers discover which teaching strategies and methods have the most positive impact on students' learning outcomes. Through regular evaluation of academic performance and classroom participation, the quality of teaching can be continuously improved to ensure that teaching activities are always adjusted around students' actual needs. This can achieve comprehensive tracking of dynamic changes in the teaching process, thereby continuously improving students' learning outcomes and sense of participation.Finally, by conducting teaching supervision and evaluation on students' knowledge learning progress and regression rates and classroom participation before and after teaching recommendations based on dynamic learning data, the effectiveness of teaching activities can be presented in a visual way. The biggest advantage of this process is that it can help teachers, managers and decision makers see the advantages and disadvantages of the teaching process more clearly through intuitive charts and data displays. The visualization results of teaching supervision and evaluation provide an important basis for improving teaching quality. In this way, schools or educational institutions can track the teaching effect of each class in real time, discover possible problems and make adjustments in time, and help teachers judge whether the teaching methods have effectively stimulated students' interest and enthusiasm in learning. Through timely visual feedback mechanism, it not only enhances the transparency and ease of use of data, but also improves the scientific nature of the teaching process and promotes the continuous improvement of education quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings: Figure 1 It is a module schematic diagram of the real-time dynamic learning data monitoring and visualization platform of the present invention; Figure 2 for Figure 1 Functional flow diagram of the teaching status connection collection and reporting module; Figure 3 for Figure 1 Schematic diagram of the functional flow of the teaching status modeling and analysis module. DETAILED DESCRIPTION

[0016] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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 technicians in this field without creative work are within the scope of protection of the present invention.

[0017] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0018] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0019] To achieve this, please refer to Figures 1 to 3 The present invention provides a real-time dynamic learning data monitoring and visualization method, the method comprising the following steps: The teaching status docking collection and reporting module is used to build a corresponding teaching basic status collection model through the combination of teaching affairs visual data docking, learning and examination affairs data docking, and teaching environment data docking with the teaching basic status data warehouse; the corresponding dynamic learning data is collected and monitored in real time and synchronously by using the docking corresponding to the teaching basic status collection model, and reported to the corresponding teaching basic status data warehouse; The teaching state modeling and analysis module is used to perform teaching modeling feature analysis on dynamic learning data in the teaching basic state data warehouse to obtain a classroom teaching visual behavior feature set, a student learning behavior feature set, and a student environmental influencing factor set; based on the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environmental influencing factor set, a personalized teaching analysis is performed on the corresponding students to generate personalized teaching suggestions for classroom students; The pre- and post-teaching recommendation analysis module is used to obtain the students' learning performance and class participation before and after the teaching recommendation based on the personalized teaching suggestions for the students in the classroom, and to analyze the students' knowledge progress and regression on their learning performance before and after the teaching recommendation to obtain the students' knowledge learning progress and regression rate before and after the teaching recommendation; The teaching supervision and evaluation visualization module is used to visualize the teaching supervision and evaluation of students' knowledge learning progress and regression rates and classroom participation before and after teaching recommendations based on dynamic learning data, so as to generate corresponding classroom teaching supervision visualization results.

[0020] In the embodiment of the present invention, please refer to Figure 1 FIG. 1 is a schematic diagram of modules of the real-time dynamic learning data monitoring and visualization system of the present invention. In this example, the real-time dynamic learning data monitoring and visualization system includes the following modules: S1: Teaching status docking collection and reporting module, used to build a corresponding teaching basic status collection model through the combination of teaching affairs visual data docking, learning and examination affairs data docking and teaching environment data docking with the teaching basic status data warehouse; use the docking corresponding to the teaching basic status collection model to monitor and collect the corresponding dynamic learning data in real time and synchronously, and report it to the corresponding teaching basic status data warehouse; In an embodiment of the present invention, by using Python to write a data docking script, the teaching visual data docking is realized through the RESTful API interface, and the teaching visual data, such as the teacher's teaching screen and the student's classroom performance, are obtained from the surveillance cameras and intelligent teaching equipment installed in the classroom. The data is transmitted to the basic teaching status data warehouse. For the learning and examination data docking, with the help of the SQL database connection tool, a connection is established with the school's teaching management system database to extract the students' course grades, attendance records, homework completion status and other learning and examination data. In terms of the teaching environment data docking, the sensor data acquisition module is used to connect the temperature and humidity sensors, light sensors, noise sensors and other equipment in the classroom to collect classroom environment data in real time. After integrating these three types of data, the data warehouse ETL (Extract, Transform, Load) tool, such as Apache Sqoop, is used to extract, transform and load the data into the basic teaching status data warehouse, and a basic teaching status acquisition model is constructed. The model automatically monitors and collects the corresponding dynamic learning data through docking in real time at a set time interval, such as every 15 minutes, and reports it to the basic teaching status data warehouse to ensure the timeliness and accuracy of the data.

[0021] S2: Teaching state modeling and analysis module, used to perform teaching modeling feature analysis on dynamic learning data in the teaching basic state data warehouse to obtain a classroom teaching visual behavior feature set, a student learning behavior feature set, and a student environmental influencing factor set; based on the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environmental influencing factor set, personalized teaching analysis is performed on the corresponding students to generate personalized teaching suggestions for classroom students; In an embodiment of the present invention, in a teaching basic status data warehouse, Python data analysis libraries such as pandas, numpy and scikit-learn are used to perform teaching modeling feature analysis. From the classroom dynamic visual data in the dynamic learning data, with the help of computer vision algorithms, such as a target detection algorithm based on a convolutional neural network, the teacher's teaching behavior (explaining, demonstrating, asking questions, etc.) and the student's classroom behavior (listening attentively, whispering, etc.) are identified to form a classroom teaching visual behavior feature set. For the learning dynamic behavior data, the number of course visits, learning time, homework submission time, etc. are calculated through a data analysis method. The concentration is evaluated in combination with the student's body movements and sitting posture to generate a student learning behavior feature set. From the teaching dynamic environment data, the influence of environmental factors such as classroom temperature, humidity, and light intensity on the student's concentration is analyzed to obtain a student environment influencing factor set. By using an association rule mining algorithm, such as an Apriori algorithm, the relationship between these feature sets and the student's learning performance is analyzed. According to the student's weak knowledge points and learning characteristics, personalized teaching suggestions for students in the classroom are generated. For example, for a student who is weak in the mathematical geometry part and is interested in animation demonstrations, relevant animation teaching videos and targeted exercises are recommended.

[0022] S3: The pre- and post-teaching recommendation analysis module is used to obtain the students' learning performance and class participation before and after the teaching recommendation based on the personalized teaching suggestions for the students in the classroom, and to analyze the students' knowledge progress and regression on their learning performance before and after the teaching recommendation to obtain the students' knowledge learning progress and regression rate before and after the teaching recommendation; In an embodiment of the present invention, teaching recommendations are implemented in the school's teaching management system based on the generated personalized teaching suggestions for classroom students. Before the recommendation, the student's academic performance in each subject is extracted from the performance management database, including regular homework performance, unit test performance and examination performance, and the comprehensive performance is calculated according to a certain weighted ratio; at the same time, the student's classroom participation data, such as the number of speeches and the length of time for group discussion participation, is obtained from the classroom interaction record system. After the teaching recommendation is implemented for a period of time, such as a semester, these data are obtained again, and a calculation script is written using Python to perform an analysis of the student's corresponding academic performance before and after the teaching recommendation. Conduct analysis. For each subject, calculate the difference between the scores before and after recommendation to obtain the space for learning progress and regression; calculate the ratio of the scores after recommendation to the scores before recommendation to obtain the amplitude of learning progress and regression. Add up the amplitudes and the ratios of the progress and regression spaces of each subject to calculate the students' corresponding knowledge learning progress and regression rates before and after the teaching recommendation. For example, a student's mathematics improvement space is 5 points, and the improvement amplitude is 1.05, and the Chinese regression space is -3 points, and the regression amplitude is 0.95. The comprehensive calculation results show that the knowledge learning progress and regression rate is -1.84, indicating that there is a certain overall regression. Finally, the students' knowledge learning progress and regression rates before and after the teaching recommendation are obtained.

[0023] S4: Teaching supervision and evaluation visualization module is used to visualize the teaching supervision and evaluation of students' knowledge learning progress and regression rates and classroom participation before and after teaching recommendations based on dynamic learning data, so as to generate corresponding classroom teaching supervision visualization results.

[0024] In an embodiment of the present invention, by using Python visualization libraries such as matplotlib and seaborn, the teaching supervision evaluation of students' corresponding knowledge learning progress and regression rates and classroom participation before and after teaching recommendations is visualized based on dynamic learning data. First, a data set containing knowledge learning progress and regression rates, classroom participation and related influencing factors (such as teacher teaching methods, teaching environment factors, etc.) is created, and a bar graph is drawn using the bar() function of matplotlib to compare the knowledge learning progress and regression rates and classroom participation of students in different classes or under different teachers. A scatter plot is drawn using the scatter() function of seaborn to analyze the relationship between teaching environment factors (such as classroom temperature) and knowledge learning progress and regression rates or classroom participation. A linear regression fitting line is added through the regplot() function to intuitively display its linear correlation. For example, it is found through visualization that within a suitable classroom temperature range, students' knowledge learning progress and regression rates and classroom participation are positively correlated. These visualization charts are sorted and summarized to finally generate corresponding classroom teaching supervision visualization results, providing intuitive data display and decision-making basis for teaching supervision.

[0025] Furthermore, the teaching status docking collection and reporting module includes the following functions: By connecting the teaching affairs visual data, learning and examination affairs data, and teaching environment data with the teaching basic status data warehouse, a corresponding teaching basic status acquisition model is constructed; By configuring the corresponding teaching status synchronization collection time point, and based on the teaching status synchronization collection time point, using the corresponding docking of the teaching basic status collection model to monitor and collect the corresponding dynamic learning data in real time, including classroom dynamic visual data, learning dynamic behavior data and teaching dynamic environment data; Stream-process the corresponding dynamic learning data to generate a basic teaching state data stream; The basic teaching status data stream is transmitted and reported to the data sub-database in the basic teaching status data warehouse corresponding to the basic teaching status acquisition model.

[0026] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 The functional flow diagram of the teaching status docking collection and reporting module in this embodiment includes the following functions: S11: Construct the corresponding basic teaching status acquisition model by connecting the teaching affairs visual data, learning and examination affairs data, and teaching environment data with the teaching basic status data warehouse; In the embodiment of the present invention, the construction of the basic state collection model of teaching relies on the combination of multiple data docking interfaces and the basic state data warehouse of teaching. For the docking of teaching visual data, a special visual data interface device is used, which can be connected with visual devices such as monitoring cameras and electronic whiteboards in the classroom to realize the collection of teaching visual data such as classroom pictures and teacher's blackboard writing; the docking of learning and examination data is connected to the school's teaching management system with the basic state collection model of teaching through a data transmission protocol to ensure that the students' test scores, attendance records and other learning and examination data can be synchronized in real time; the docking of teaching environment data uses the temperature and humidity sensors, light sensors and other devices installed in the classroom to collect the teaching environment data such as the temperature, humidity and light intensity of the classroom. The data obtained by these dockings are integrated with the basic state data warehouse of teaching, and the database management system (such as MySQL) is used for data storage and management to construct the corresponding basic state collection model of teaching.

[0027] S12: by configuring the corresponding teaching state synchronous collection time point, and based on the teaching state synchronous collection time point, using the corresponding docking of the teaching basic state collection model to synchronously monitor and collect the corresponding dynamic learning data in real time, including classroom dynamic visual data, learning dynamic behavior data and teaching dynamic environment data; In an embodiment of the present invention, in order to realize the real-time synchronous monitoring and collection of dynamic learning data, it is necessary to configure the teaching status synchronous collection time point, so as to set multiple collection time points every day according to the school's course schedule and teaching needs by using a timed task scheduler (such as Cron in the Linux system), for example, data is collected at the beginning, middle and end of each class. Based on these time points, the teaching basic status collection model starts working. For classroom dynamic visual data, the monitoring camera captures the picture and records the video at the set time point; the learning dynamic behavior data is collected in real time by recording the students' operation records on the learning platform (such as login time, learning time, homework submission status, etc.); the teaching dynamic environment data is collected by sensors at set time intervals, and these data are synchronized to the teaching basic status collection model in real time to ensure that comprehensive dynamic learning data is collected.

[0028] S13: Stream-processing the corresponding dynamic learning data to generate a teaching basic state data stream; In an embodiment of the present invention, the collected dynamic learning data is stream processed so as to process the data by using a streaming processing framework (such as Apache Kafka and Apache Flink). First, the dynamic learning data is sent to the Kafka message queue. Kafka is responsible for the storage and distribution of the data. Flink, as a data processing engine, reads the data from Kafka and performs cleaning, conversion and aggregation operations on the data. For example, image recognition is performed on the dynamic visual data of the classroom to extract the expression and action information of the students; statistical analysis is performed on the dynamic learning behavior data to calculate the learning activity of the students; the dynamic teaching environment data is standardized to have a unified format. The processed data forms a basic teaching status data stream, which provides a basis for subsequent data analysis and application.

[0029] S14: The basic teaching status data stream is transmitted and reported to the data sub-database in the basic teaching status data warehouse corresponding to the basic teaching status acquisition model.

[0030] In an embodiment of the present invention, the teaching basic status data stream is transmitted and reported to the data sub-library in the teaching basic status data warehouse corresponding to the teaching basic status acquisition model, so that the streaming teaching basic status data stream is transmitted from Flink to the data sub-library in the MySQL database by using a data transmission tool (such as Sqoop). During the transmission process, the integrity and accuracy of the data are ensured, and the transmitted data is checked by a data verification mechanism to prevent data loss or error. The data sub-library is classified and stored according to the type and purpose of the data, for example, classroom dynamic visual data is stored in the image and video sub-library, learning dynamic behavior data is stored in the learning record sub-library, and teaching dynamic environment data is stored in the environment monitoring sub-library. In this way, the teaching basic status data warehouse can effectively manage and store dynamic learning data, and provide support for subsequent teaching analysis and decision-making.

[0031] Furthermore, the real-time synchronous monitoring and acquisition of the corresponding dynamic learning data based on the teaching state synchronous acquisition time point using the teaching basic state acquisition model corresponding to the docking includes: Based on the teaching status synchronous collection time point, the teaching affairs visual data corresponding to the basic teaching status collection model is used to connect the real-time monitoring collection of the corresponding classroom dynamic visual data, including the classroom panorama, teacher teaching activities and student classroom performance corresponding video stream data; In the embodiment of the present invention, at the time point of synchronous collection of teaching status, the teaching affairs visual data corresponding to the teaching basic status collection model is used to connect to the real-time monitoring and collection of classroom dynamic visual data. A high-definition network camera is installed at the four corners of the classroom ceiling, and the camera is connected to the network switch through an RJ45 network cable, and then connected to the school local area network to achieve the connection with the teaching affairs visual data of the teaching basic status collection model. The camera adopts a 360-degree panoramic shooting mode to obtain panoramic video stream data of the classroom, which can fully present the layout of the classroom and the location distribution of students and teachers. A camera specially designed to capture the teacher's teaching activities is installed above the podium. The camera has automatic The dynamic focus and low-light shooting functions ensure that the video stream data of teachers' teaching activities such as blackboard writing, teaching aids display, and body movements can be clearly recorded under different lighting conditions. A small camera is set in front of each student seat in the classroom to capture only the upper body of the student, which is used to collect video stream data corresponding to the student's classroom performance, such as facial expressions, hand-raising and speaking movements, and concentration. These cameras synchronize the collection time points according to the teaching status, such as 10 minutes after the start of each class, halfway through the course, and 10 minutes before the end of the class. The shooting is automatically started and the video stream data is transmitted in real time through the network to the storage server corresponding to the basic teaching status collection model.

[0032] Preferably, based on the teaching status synchronous collection time point, the learning examination data corresponding to the teaching basic status collection model is connected to the real-time monitoring collection of the corresponding learning dynamic behavior data, including the number of course visits, learning time, homework submission time, online test scores, students' corresponding body movements and sitting status in class; In the embodiment of the present invention, by synchronously collecting the time point based on the teaching status, the learning dynamic behavior data is collected in real time through the learning examination data corresponding to the teaching basic status collection model, and a data collection plug-in is embedded in the school's learning management system (such as the Moodle platform). The plug-in establishes a data connection with the teaching basic status collection model. The learning management system records the number of students' course visits in real time. Whenever a student logs in to the system to visit the course page, the plug-in automatically sends the visit record to the teaching basic status collection model, and uses the system's own timer function to record the student's learning time after each login. At the teaching status synchronization collection time point, such as 10 p.m. every day At the same time, the plug-in synchronizes the learning time data of all students on that day to the collection model. For the homework submission time, when students submit homework in the learning management system, the system automatically records the submission time and uploads it in real time by the plug-in. After the students complete the test and submit the answers, the system automatically corrects the online test scores and transmits the scores to the collection model through the plug-in. Motion capture equipment is installed in the classroom, such as a body motion capture camera based on infrared technology. At the teaching state synchronization collection time point, the camera captures the corresponding body motion data of students in the classroom, such as walking, standing, writing, sitting twisting, etc., and sends these data to the teaching basic state collection model through the wireless transmission module. At the same time, pressure sensors and posture recognition devices are installed on the students' seats to monitor the students' sitting status in real time, and transmit the data to the collection model at the collection time point to fully obtain the dynamic learning behavior data.

[0033] Preferably, based on the teaching status synchronization collection time point, the teaching environment data corresponding to the teaching basic status collection model is used to connect the real-time monitoring and collection of corresponding teaching dynamic environment data, including the teacher's teaching content, students' questions and answers, classroom discussion sounds corresponding to the voice text data and classroom learning environment data. The classroom learning environment data includes the corresponding temperature, humidity and light intensity of the classroom.

[0034] In an embodiment of the present invention, by synchronously collecting the time points according to the teaching status, the teaching environment data corresponding to the teaching basic status collection model is connected to monitor and collect the teaching dynamic environment data in real time, and multiple microphones are installed on the walls of the classroom. The microphones are connected to the audio collector through audio cables. The audio collector is connected to the teaching basic status collection model. The microphones convert the sound signals such as the teacher's teaching content, students' questions and answers, and classroom discussion sounds into electrical signals. After being digitally processed by the audio collector, they are converted into voice text data using voice recognition software (such as iFlytek voice recognition engine). At the teaching status synchronous collection time points, such as after each class, the class content of the class will be recorded. The voice and text data are transmitted to the basic teaching status acquisition model. Temperature and humidity sensors (such as DHT11 temperature and humidity sensors) and light intensity sensors (such as BH1750 light sensors) are installed in the classroom. These sensors are connected to the data acquisition module through the I2C communication protocol. The data acquisition module is then connected to the basic teaching status acquisition model. The sensor collects the corresponding temperature, humidity and light intensity data of the classroom every 10 minutes. At the teaching status synchronization acquisition time point, such as 5 pm every day, the data acquisition module summarizes the classroom learning environment data collected that day and transmits it to the basic teaching status acquisition model to realize the comprehensive collection of dynamic teaching environment data.

[0035] Furthermore, the teaching status modeling and analysis module includes the following functions: By combining convolutional neural networks and recurrent neural networks in the teaching basic status data warehouse to build a corresponding classroom visual behavior analysis model, and obtaining the corresponding classroom teacher and student behavior annotation data, the classroom visual behavior analysis model is trained based on the classroom teacher and student behavior annotation data to generate a classroom behavior analysis model that can identify the corresponding teacher and student behaviors; Input the classroom dynamic visual data into the classroom behavior analysis model that can identify the corresponding teacher and student behaviors to perform classroom visual behavior feature analysis, so as to identify the corresponding teaching behavior features of teachers such as explanation, demonstration and questioning, and the corresponding classroom behavior features of students such as attentive listening, whispering, discussion and absent-mindedness, and obtain the classroom teaching visual behavior feature set; The spatiotemporal convolutional network is used to analyze the spatiotemporal variation characteristics of the course access times, learning time, homework submission time, and online test scores corresponding to the learning dynamic behavior data, so as to analyze the variation characteristics of the corresponding student behaviors in the time and space sequence, and the student concentration statistics are performed on the corresponding body movements and sitting postures of the students in the classroom in the learning dynamic behavior data to obtain the corresponding concentration of the students in the classroom. At the same time, the variation characteristics of the student behaviors in the time and space sequence and the corresponding concentration of the students in the classroom are combined to obtain the student learning behavior feature set; Based on the dynamic teaching environment data, the learning impact assessment of students' corresponding concentration in the classroom is carried out to determine the environmental factors that affect students' concentration during the teacher's classroom teaching process, so as to obtain the student environmental influence factor set; key mining and dimensionality reduction processing are carried out between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influence factor set, so as to generate the dimensionality reduction feature set of the basic teaching state; Through the dimensionality reduction feature set analysis of the basic teaching status, the corresponding classroom learning evaluation of the students is obtained, including the corresponding learning history, interest preferences and knowledge mastery of the students. Personalized teaching analysis is performed based on the corresponding classroom learning evaluation of the students, so as to recommend corresponding learning resources and learning paths for the students' weak knowledge points and generate personalized teaching suggestions for students in the classroom.

[0036] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 The functional flow diagram of the teaching status modeling and analysis module in the present embodiment includes the following functions: S21: By combining the convolutional neural network and the recurrent neural network in the teaching basic status data warehouse to build a corresponding classroom visual behavior analysis model, and obtaining the corresponding classroom teacher and student behavior annotation data, the classroom visual behavior analysis model is trained based on the classroom teacher and student behavior annotation data to generate a classroom behavior analysis model that can identify the corresponding teacher and student behaviors; In an embodiment of the present invention, a classroom visual behavior analysis model is constructed by using Python's deep learning framework TensorFlow in a teaching basic status data warehouse, and the advantages of convolutional neural network (CNN) processing image spatial features and recurrent neural network (RNN) processing time series features are combined to design a model structure. First, a labeling tool (such as LabelImg) is used to label the classroom dynamic visual data obtained from the teaching basic status acquisition model, marking the teacher's explanation, demonstration, questioning and other behaviors as well as the students' attentive listening, whispering, discussion, and slacking behaviors to form classroom teacher and student behavior labeling data. The labeled data is divided into 70% training set, 20% validation set, and 10% test set ratios. The model is built in TensorFlow, and the input layer receives an image sequence of classroom dynamic visual data. The CNN layer extracts spatial features in the image through a convolution kernel, such as teacher body movements, student facial expressions and other features. The RNN layer processes time series, captures changes in behavior over time, and trains the model by using the cross-entropy loss function and the Adam optimizer. The model parameters are continuously adjusted during the training process. After multiple rounds of training, the model performance is evaluated on the validation set. When indicators such as the model accuracy and recall rate reach the preset standards, a classroom behavior analysis model that can identify the corresponding teacher and student behaviors is finally generated.

[0037] S22: inputting classroom dynamic visual data into a classroom behavior analysis model that can identify teacher and student behaviors to perform classroom visual behavior feature analysis, so as to identify the teacher's corresponding teaching behavior features of explanation, demonstration and questioning, and identify the students' corresponding classroom behavior features of listening attentively, whispering, discussing and being absent-minded, and obtain a classroom teaching visual behavior feature set; In an embodiment of the present invention, classroom dynamic visual data is divided into image sequences in chronological order and input into a trained classroom behavior analysis model. The CNN layer of the model first extracts features from the input image and identifies spatial features such as body parts and postures of teachers and students. The RNN layer then processes the changes in these features in the time series to identify the corresponding teaching behavior features of the teacher's explanation, demonstration and questioning. For example, when the model detects that the teacher holds teaching aids and performs a display action sequence, it is determined to be a demonstration behavior; when the teacher opens his mouth to speak to the students and is accompanied by a certain sequence of gesture actions, it is determined to be an explanation behavior. For student behavior, when the model recognizes an image sequence of students' faces facing the blackboard and their expressions are focused, it is determined to be attentive listening; when an image sequence of two students' heads close together and accompanied by frequent body movements is detected, it is determined to be whispering. By analyzing the entire section of classroom dynamic visual data, a set of classroom teaching visual behavior features is finally obtained, recording information such as the time of occurrence and duration of each behavior.

[0038] S23: Use a spatiotemporal convolutional network to analyze the spatiotemporal variation characteristics of the course access times, learning time, homework submission time, and online test scores corresponding to the learning dynamic behavior data, so as to analyze the variation characteristics of the corresponding student behaviors in the time and space sequence, and perform student concentration statistics on the corresponding body movements and sitting postures of the students in the classroom in the learning dynamic behavior data, so as to obtain the corresponding concentration of the students in the classroom, and at the same time, merge the variation characteristics of the student behaviors in the time and space sequence and the corresponding concentration of the students in the classroom to obtain the student learning behavior feature set; In an embodiment of the present invention, the spatiotemporal change characteristics of the course access times, learning time, homework submission time and online test scores in the learning dynamic behavior data are analyzed by using the spatiotemporal convolutional network (STCN) in the PyTorch framework of Python, so that these data are sorted into spatiotemporal sequence data according to student ID and time sequence and input into the STCN model. The spatiotemporal convolution layer of the STCN performs convolution operations on the data in the time and space dimensions to extract the corresponding change characteristics of the student behavior in the time and space sequence. For example, the increase and decrease trend of the number of course visits of a certain student in a specific time period, the change law of the learning time and the concentrated time period of homework submission time are analyzed. For the corresponding body movements and sitting states of the students in the classroom, the posture is estimated by a pre-trained deep learning-based posture recognition model (such as OpenPose), and the ratio between the stability of the student's good sitting posture and the degree of center of gravity deviation is counted to obtain the corresponding concentration of the students in the classroom. The corresponding change characteristics of the student behavior in the time and space sequence and the corresponding concentration of the student in the classroom are merged to form a student learning behavior feature set, which provides comprehensive data support for subsequent analysis.

[0039] S24: Based on the teaching dynamic environment data, the learning impact assessment is conducted on the students' corresponding concentration in the classroom to determine the environmental factors that affect the students' concentration during the teacher's classroom teaching process, so as to obtain the student environmental influence factor set; key mining and dimensionality reduction processing are performed between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influence factor set to generate the dimensionality reduction feature set of the basic teaching state; In an embodiment of the present invention, based on the teaching dynamic environment data, including the voice text data corresponding to the teacher's teaching content, students' questions and answers, classroom discussion sounds, and classroom learning environment data (temperature, humidity, light intensity), the impact on the students' corresponding concentration in the classroom is analyzed, and the natural language processing technology (NLP) is used to perform sentiment analysis, theme extraction, etc. on the voice text data to judge the attractiveness of the teacher's teaching content, the activeness of the classroom discussion, etc. on the student's concentration. At the same time, through data analysis methods, such as correlation analysis, the relationship between classroom temperature, humidity, light intensity and student concentration is studied. For example, it is found that when the temperature is too high, the student's concentration decreases. The environmental factors that affect the student's concentration are sorted into a student environment influencing factor set. For the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environment influencing factor set, the principal component analysis (PCA) and other dimensionality reduction algorithms are used for key mining and dimensionality reduction processing. The PCA algorithm converts high-dimensional data into low-dimensional data through linear transformation, retains the main features of the data, removes redundant information, and finally generates a teaching basic state dimensionality reduction feature set for subsequent efficient analysis.

[0040] S25: Obtain the students’ corresponding classroom learning evaluations through the dimensionality reduction feature set analysis of the basic teaching status, including the students’ corresponding learning history, interest preferences, and knowledge mastery status, and conduct personalized teaching analysis based on the students’ corresponding classroom learning evaluations, so as to recommend corresponding learning resources and learning paths for the students’ corresponding weak knowledge points, and generate personalized teaching suggestions for students in the classroom.

[0041] In an embodiment of the present invention, by reducing the dimension feature set according to the basic teaching state, using data mining and machine learning algorithm analysis to obtain the corresponding classroom learning evaluation of the student, so as to analyze the student's learning history through an association rule mining algorithm (such as an Apriori algorithm) to find out the students' learning performance rules in different courses and different time periods, and use a clustering algorithm (such as K-Means clustering) to analyze the students' interest preferences, cluster the students according to the characteristics such as the students' participation in different course contents and the length of study, and determine their interest directions, and by constructing a knowledge mastery evaluation model (such as a student knowledge mastery prediction model based on deep learning), the students' knowledge mastery status is evaluated according to data such as the students' homework scores and online test scores. Based on these classroom learning evaluations, for the students' weak knowledge points, relevant teaching videos, exercises and other learning resources are screened from the school's learning resource library, and a path planning algorithm (such as a Dijkstra algorithm) is used to plan personalized learning paths for students according to their learning progress, interest preferences and knowledge mastery status, and generate personalized teaching suggestions for students in the classroom. For example, geometry-related teaching videos are recommended to students who have weak mathematics scores and are interested in graphic knowledge, and a learning path from basic concept learning to advanced exercises is planned.

[0042] Furthermore, the student concentration statistics of the corresponding body movements and sitting postures of the students in the classroom in the learning dynamic behavior data include: Performing time-series synchronization processing on the corresponding body movements and sitting postures of students in the classroom in the learning dynamic behavior data to generate corresponding student body movement sequences and student sitting posture sequences within the same time-series range; In an embodiment of the present invention, the corresponding body movements and sitting status data of students in the classroom in the learning dynamic behavior data are processed in the teaching basic status data warehouse. It is assumed that the body movement data and the sitting status data are stored in different time series data tables, and each data record has a timestamp, so as to read these data by using Python's pandas library, sort the data by timestamp, and determine a unified time series range according to the start time and end time of the class. For example, a class starts at 9 am and ends at 9:45 am. For the body movement data, all records in the time period from 9 am to 9:45 am are screened out, and rearranged in chronological order to generate a student body movement sequence. Similarly, the sitting status data is screened and sorted to generate a corresponding student sitting status sequence within the same time series range. For example, the body movement data records the student's hand raising at 9:10 am, turning around at 9:20 am, and the actions are sorted into sequences in order; the sitting status data records the sitting posture at 9:15 am, the body leaning forward at 9:30 am, and the corresponding sequences are also sorted.

[0043] Preferably, the corresponding student head rotation angle and student arm swing angle are obtained through the student body movement sequence, and the center of gravity shift analysis is performed according to the student head rotation angle and the student arm swing angle to obtain the degree of student center of gravity shift; In an embodiment of the present invention, computer vision technology is used to analyze the student's body movement sequence with the help of the OpenCV library to obtain the corresponding student head rotation angle and student arm swing angle. For the student's head rotation angle, a pre-trained deep learning-based head detection model (such as a Haar cascade detector) is first used to detect the student's head position in each frame image of the body movement sequence. Then, according to the position change of the head in consecutive frames, the head rotation angle is calculated using trigonometric functions. For example, at a certain moment, the center coordinates of the head are (x1, y1), and at the next moment they are (x2, y2). By calculating The angle between the two points is used to obtain the head rotation angle. For the student's arm swing angle, the skeleton key point detection algorithm (such as the arm skeleton key point detection in OpenPose) is used to determine the position of the arm at different times, and then the arm swing angle is calculated. According to the head rotation angle and the arm swing angle, combined with the human body center of gravity calculation formula (assuming that the human body center of gravity is related to the position of the head and arms, the weight coefficient can be obtained through experiments or theoretical deduction), the center of gravity offset analysis is performed. For example, assuming that the head position weight is 0.6 and the arm position weight is 0.4, the degree of student center of gravity offset is obtained by calculation, such as the center of gravity offset of 5 cm in a certain period of time.

[0044] Preferably, the duration of the student's gaze at the blackboard is obtained through the student's body movement sequence, and the number of twists of the sitting body is obtained through the student's sitting state sequence; In an embodiment of the present invention, the student's body movement sequence is processed by continuing to use the OpenCV library to obtain the corresponding student's gaze duration of the blackboard, and the direction of the student's eyes is detected in the image of the body movement sequence, and the position relationship of the eye feature points (such as the corner of the eye, the center of the eyeball) is used to determine whether the student is looking at the blackboard. When it is detected that the student's eyes are facing the blackboard, the timing starts. If the eyes continue to face the blackboard in subsequent consecutive frames, the timing continues until the eye direction changes. The duration of each gaze at the blackboard is accumulated to obtain the student's gaze duration. For example, in a class, the student gazes at the blackboard three times, which lasts for 10 minutes, 8 minutes and 5 minutes respectively, so the total gaze duration is 23 minutes. For the number of sitting body twists, in the student's sitting state sequence, the body posture changes are counted. For example, a threshold value of body posture change is pre-set, and when the difference in body posture between two adjacent frames of the sitting state sequence exceeds the threshold, it is determined to be a body twist, and the number of sitting body twists in a class is counted, assuming that it is 15 times.

[0045] Preferably, the sitting posture stability evaluation is calculated based on the duration of the student's gaze at the blackboard and the number of twists of the sitting body, so as to obtain the student's sitting posture stability; In an embodiment of the present invention, a sitting posture stability evaluation is calculated based on the duration of the student's gaze at the blackboard and the number of times the student twists his body while sitting, so as to obtain the student's sitting posture stability, and a sitting posture stability evaluation formula is set, for example, sitting posture stability = duration of the student's gaze at the blackboard ÷ (duration of the student's gaze at the blackboard + number of times the student twists his body while sitting × duration of impact of each twist), assuming that the duration of impact of each twist is set to 1 minute based on experience, in the above example, the student stares at the blackboard for 23 minutes, and the number of times the student twists his body while sitting is 15 times, then the sitting posture stability = 23 ÷ (23 + 15 × 1) ≈ 0.61, the closer the value is to 1, the more stable the student's sitting posture, and the smaller the value is, the worse the sitting posture stability. Through this calculation result, the student's sitting posture stability in class can be intuitively understood, providing a basis for subsequent concentration statistics.

[0046] Preferably, the student concentration is counted according to the degree of the student's center of gravity shift and the stability of the student's sitting posture to obtain the student's corresponding concentration in class.

[0047] In an embodiment of the present invention, by statistically calculating the ratio between the previously calculated degree of stability of the student's sitting posture and the degree of deviation of the student's center of gravity, the corresponding degree of concentration of the student in class is finally obtained.

[0048] Furthermore, the learning impact assessment and judgment of the students' corresponding concentration in class based on the teaching dynamic environment data includes: Based on the various environmental factors in the teaching dynamic environment data, a linear correlation impact assessment is performed on the corresponding concentration of students in the classroom to obtain the linear correlation influence coefficient between each teaching environment factor and concentration; In an embodiment of the present invention, a linear correlation impact assessment is performed on each environmental factor in the teaching dynamic environment data and the corresponding concentration of the students in the classroom by using Python's pandas and numpy libraries. It is assumed that the teaching dynamic environment data is stored in a pandas DataFrame object, including classroom temperature, humidity, light intensity, emotional tendency of the teacher's teaching content (obtained by natural language processing), classroom discussion activity (based on voice text data statistics) and other environmental factor columns, and also includes a student concentration column, and the Pearson correlation coefficient between each environmental factor column and the concentration column is calculated by using the pandas.DataFrame.corr() method. The method automatically calculates the linear correlation between the two columns of data. For example, the correlation coefficient between the classroom temperature and the concentration is calculated to be -0.3, indicating that when the temperature rises, the student's concentration will tend to decrease; the correlation coefficient between the emotional tendency of the teacher's teaching content and the concentration is 0.5, which means that the more positive the teaching content is, the higher the student's concentration is. In this way, the linear correlation influence coefficients between each teaching environment factor and the concentration are obtained, and these coefficients quantify the linear influence of each environmental factor on the student's concentration.

[0049] Preferably, based on the linear correlation influence coefficient between each teaching environment factor and the concentration, each environmental factor in the teaching dynamic environment data is screened for learning impact judgment. If the linear correlation influence coefficient between the corresponding teaching environment factor and the concentration is positively correlated, then the corresponding teaching environment factor is the environmental factor that affects the student's concentration; if the linear correlation influence coefficient between the corresponding teaching environment factor and the concentration is negatively correlated, then the corresponding teaching environment factor is not the environmental factor that affects the student's concentration, thereby judging and screening out the environmental factors that affect the student's concentration during the teacher's classroom teaching process to obtain a set of student environment influencing factors.

[0050] In an embodiment of the present invention, by using the linear correlation influence coefficient between each teaching environment factor and concentration obtained previously, each environmental factor in the teaching dynamic environment data is screened for learning impact judgment, so as to traverse the linear correlation influence coefficients of all environmental factors, and for each coefficient, its positivity is judged. For example, the correlation coefficient between classroom humidity and concentration is -0.1. Since it is negatively correlated, classroom humidity is not an environmental factor that affects students' concentration; and the correlation coefficient between classroom discussion activity and concentration is 0.4, which is positively correlated. Then classroom discussion activity is the environmental factor that affects students' concentration. All teaching environment factors that are positively correlated are screened out and organized into a new data set, namely, a set of student environment influencing factors. This set clarifies which environmental factors have a positive promoting effect on students' concentration during the teacher's classroom teaching process, and provides a key basis for subsequent optimization of the teaching environment and improvement of teaching quality.

[0051] Furthermore, the key mining and dimensionality reduction processing between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influencing factor set includes: The correlation measurement calculation is performed between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influencing factor set, so as to quantitatively calculate the correlation coefficient between the two sub-features and form a corresponding correlation matrix to generate the basic state feature correlation matrix of teaching; In an embodiment of the present invention, by using Python's pandas and numpy libraries, relevant metric calculations are performed on the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environment influencing factor set. It is assumed that the classroom teaching visual behavior feature set is stored in a pandas DataFrame object visual_df, including sub-feature columns such as the teacher's explanation time and the number of students whispering; the student learning behavior feature set is stored in learning_df, including sub-feature columns such as the number of course visits and concentration; the student environment influencing factor set is stored in environment_df, including sub-feature columns such as classroom discussion activity and classroom temperature. The DataFrame objects are merged into a new DataFrameall_features_df by column, and the Pearson correlation coefficient between all sub-features is calculated by using the all_features_df.corr() method. This method automatically calculates the correlation measurement for every two sub-features. For example, the correlation coefficient between the teacher's explanation time and the student's concentration is calculated to be 0.6, and the correlation coefficient between the number of course visits and the activeness of classroom discussion is 0.8. These correlation coefficients are organized into a two-dimensional matrix, and the corresponding correlation matrix is ​​constructed. Finally, the basic teaching status feature correlation matrix is ​​generated. The rows and columns of the matrix correspond to each sub-feature, and the matrix elements are the correlation coefficients between the two sub-features.

[0052] Preferably, based on the teaching basic state feature correlation matrix, the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influencing factor set are subjected to key mining and screening of behavioral features. If the corresponding correlation coefficient between the two sub-features in the teaching basic state feature correlation matrix is ​​greater than 0.75, it is considered that the two sub-features are all key features; if the corresponding correlation coefficient between the two sub-features in the teaching basic state feature correlation matrix is ​​equal to 0.75, it is considered that there is a key feature in the two sub-features; if the corresponding correlation coefficient between the two sub-features in the teaching basic state feature correlation matrix is ​​less than 0.75, it is considered that the two sub-features are not key features; the corresponding key features are screened out according to the above judgment rules to generate the teaching basic state key feature set; In the embodiment of the present invention, based on the teaching basic state feature correlation matrix, the Python loop structure is used to perform key mining and screening of behavioral features between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environmental influencing factor set, so as to traverse each element of the teaching basic state feature correlation matrix, and for each group of correlation coefficients corresponding to each pair of sub-features, a judgment is made according to the judgment rule. For example, for the group of sub-features of teacher demonstration time and classroom discussion activity, the correlation coefficient is - 0.1, less than 0.75, then it is considered that these two sub-features are not key features; for the group of classroom discussion activity and student concentration, the correlation coefficient is 0.8, which is greater than 0.75, then both classroom discussion activity and student concentration are key features; assuming that there is a set of sub-features with a correlation coefficient of exactly 0.75, such as the number of questions asked by teachers and the enthusiasm of students in answering questions. Further, by comparing the correlation between these two sub-features and other key features or referring to expert opinions, one of them is determined to be the key feature, and all sub-features that meet the key feature conditions are screened out and stored in a new list or DataFrame to generate a key feature set of basic teaching status. These key features are of great significance for understanding the teaching process and student learning status.

[0053] Preferably, principal component dimensionality reduction processing is performed on the corresponding sub-features in the key feature set of the basic teaching state to generate a dimensionality reduction feature set of the basic teaching state.

[0054] In an embodiment of the present invention, the principal component analysis (PCA) module in the scikit-learn library of Python is used to perform principal component dimensionality reduction processing on the corresponding sub-features in the key feature set of the basic teaching state. It is assumed that the key feature set of the basic teaching state is stored in a pandas DataFrame object key_features_df. First, the PCA object is instantiated, for example, pca = PCA(n_components = 0.95). Here, 95% of the information is set to be retained, so that key_features_df is passed as input data to the fit_transform() method of the PCA object. This method performs principal component analysis on the data and converts the original high-dimensional data into low-dimensional data. For example, the original key feature set contains 10 sub-features. After PCA processing, the data is converted into 3-4 principal components, which can retain the information of the original data to the greatest extent. The data after dimensionality reduction is organized into a new DataFrame to generate a dimensionality reduction feature set of the basic teaching state. Through dimensionality reduction processing, the dimension of the data is reduced, the complexity of data processing is reduced, and key information is retained, providing a more efficient data basis for subsequent data analysis and visualization.

[0055] Furthermore, the teaching recommendation before and after analysis module includes the following functions: Based on the personalized teaching suggestions for students in class, the corresponding classroom teaching recommendations are implemented, and the corresponding learning performance and classroom participation of students before and after the teaching recommendations are obtained, where the learning performance includes the learning performance corresponding to each subject knowledge; In an embodiment of the present invention, classroom teaching recommendations are performed for students based on the generated personalized teaching recommendations for students in the classroom in the school's teaching management system. It is assumed that the personalized teaching recommendations are stored in a document, which lists in detail the recommended learning resources (such as specific teaching videos, online exercise links) and learning paths (such as learning a certain chapter of knowledge first, and then doing related exercises) for each student's weak knowledge points. The teaching management system pushes these recommendation information to students through student accounts. Students start learning after receiving the recommendations. Before the teaching recommendation, the school's grade management database is used to extract the student's learning grades corresponding to each subject knowledge, such as Chinese, mathematics, English and other subjects. The grades are calculated based on the regular homework grades, test grades, and examination grades according to a certain ratio. At the same time, the classroom interaction recording system is used to count the student's classroom participation, such as the number of speeches, the length of time for group discussion participation, etc. After the teaching recommendation is implemented for a period of time (such as one month), the student's learning grades are obtained from the grade management database again, and the classroom participation data is updated from the classroom interaction recording system to compare and analyze the effect of the teaching recommendation.

[0056] Preferably, a learning progress and regression space analysis is performed on each subject knowledge in the student's learning scores before and after the teaching recommendation, so as to quantitatively calculate the difference in the learning scores of each subject knowledge before and after the teaching recommendation, and obtain the learning progress and regression space corresponding to each subject knowledge; In an embodiment of the present invention, the learning progress and regression space of each subject knowledge in the student's learning performance before and after the teaching recommendation is analyzed by using Python's pandas library, so that the learning performance data before the teaching recommendation and the learning performance data after the teaching recommendation are read into two pandas DataFrame objects respectively, assuming that they are before_df and after_df respectively, the columns of each DataFrame correspond to different subject knowledge, and the rows correspond to different students. The operation of after_df -before_df is used to calculate the difference in the corresponding learning performance of each subject knowledge before and after the teaching recommendation column by column. For example, for a certain student, the score of mathematics before the teaching recommendation is 80 points, and the score after the teaching recommendation is 85 points. Then the learning progress and regression space of mathematics is 85-80=5 points. In this way, the learning progress and regression space corresponding to each subject knowledge is obtained. Positive numbers represent improvement space, and negative numbers represent regression space, so as to provide quantitative data for evaluating the impact of teaching recommendations on the learning performance of each subject knowledge.

[0057] Preferably, the learning progress and regression range of each subject knowledge in the student's learning scores before and after the teaching recommendation is calculated, so as to quantitatively calculate the ratio of the learning scores of each subject knowledge before and after the teaching recommendation, and obtain the learning progress and regression range corresponding to each subject knowledge; In an embodiment of the present invention, the learning progress and regression of each subject knowledge in the student's learning performance before and after the teaching recommendation is calculated by also using Python's pandas library. Based on the previously read before_df and after_df, the after_df÷before_df operation is used to calculate the ratio of the corresponding learning performance of each subject knowledge before and after the teaching recommendation. For example, a student's Chinese score before the teaching recommendation is 70 points, and the score after the teaching recommendation is 77 points. Then the Chinese learning progress and regression is 77÷70=1.1. If the ratio is greater than 1, it indicates the improvement of the performance. The larger the ratio, the more obvious the improvement; if the ratio is less than 1, it indicates the regression of the performance. Through this calculation, the learning progress and regression corresponding to each subject knowledge is obtained, which more intuitively reflects the degree to which the teaching recommendation improves or decreases the learning performance of each subject knowledge.

[0058] Preferably, the students' knowledge progress and regression are summarized and calculated based on the learning progress and regression space and the learning progress and regression amplitude corresponding to each subject knowledge, so as to obtain the students' knowledge learning progress and regression rates before and after the teaching recommendation.

[0059] In an embodiment of the present invention, the ratio of the learning progress and regression amplitude corresponding to each subject knowledge to the learning progress and regression space is calculated, and the ratios corresponding to all subject knowledge are added together to obtain the corresponding learning progress and regression rate, and finally the corresponding knowledge learning progress and regression rate of the students before and after the teaching recommendation is obtained.

[0060] Furthermore, the teaching supervision and evaluation visualization module includes the following functions: According to the classroom dynamic visual data, learning dynamic behavior data and teaching dynamic environment data corresponding to the dynamic learning data, a teaching supervision evaluation index system is constructed to generate corresponding teacher teaching quality evaluation indicators, student learning effect evaluation indicators and teaching environment evaluation indicators. The teacher teaching quality evaluation indicators include the effectiveness of teaching methods, classroom interaction effects and the improvement of student learning outcomes. The student learning effect evaluation indicators include academic performance and learning interest cultivation. The teaching environment evaluation indicators include classroom learning impact, classroom facility applicability and environmental learning atmosphere. In an embodiment of the present invention, a teaching supervision evaluation index system is constructed by using Python's pandas library and professional education evaluation theory. The teacher's teaching methods, such as the frequency and effect of explanations, demonstrations, questions, etc., are analyzed from the dynamic visual data of the classroom. If the teacher's explanation is clear and the questions can effectively guide students to think, the teaching method effectiveness score is high; the number of classroom interactions and student participation are counted to evaluate the classroom interaction effect. For the improvement of students' learning outcomes, the changes in students' academic performance and knowledge mastery over a period of time are compared to score, and the academic performance is obtained from the learning dynamic behavior data, and the comprehensive calculation is performed according to different subjects and different assessment types (homework, quizzes, exams). points; evaluate the cultivation of learning interest through the number of students' visits to the course and the enthusiasm of participating in discussions. For the teaching environment evaluation indicators, evaluate the impact of classroom learning from the degree of influence of factors such as classroom temperature, humidity, light intensity and noise in the teaching dynamic environment data on students' learning; check whether the classroom facilities (desks, chairs, blackboards, multimedia equipment, etc.) meet the teaching needs. The higher the degree of satisfaction, the higher the suitability score of the classroom facilities; judge the environmental learning atmosphere based on the sound of classroom discussions and the activeness of students' questions and answers. Organize these evaluation results into corresponding teacher teaching quality evaluation indicators, student learning effect evaluation indicators and teaching environment evaluation indicators to form a teaching supervision evaluation indicator system.

[0061] Preferably, a teaching supervision data report is established based on the scores corresponding to the teacher teaching quality evaluation index, the student learning effect evaluation index and the teaching environment evaluation index, and the students' corresponding knowledge learning progress and regression rate and classroom participation before and after the teaching recommendation, so as to generate a teacher teaching supervision quality data report; In an embodiment of the present invention, a teaching supervision data report is established by using Python's pandas library. First, an empty DataFrame object is created for storing data. For the teacher teaching quality evaluation index, the scores of the three indicators of teaching method effectiveness, classroom interaction effect, and student learning achievement improvement are filled in the corresponding columns of the DataFrame. For example, a teacher's teaching method effectiveness score is 8 points (out of 10 points), the classroom interaction effect score is 7 points, and the student learning achievement improvement score is 7.5 points. For the student learning effect evaluation index, the scores of academic performance and learning interest cultivation are filled in. Assuming that the comprehensive score of academic performance is 85 points (out of 100 points), the score of classroom interaction effect is 7 points, and the score of student learning achievement improvement is 7.5 points. For the teaching environment evaluation indicators, fill in the scores of classroom learning impact, classroom facility applicability, and environmental learning atmosphere. For example, the classroom learning impact score is 8 points, the classroom facility applicability score is 9 points, and the environmental learning atmosphere score is 7 points. At the same time, add two columns to the DataFrame to record the students' knowledge learning progress and regression rates and classroom participation data before and after the teaching recommendation. For example, the knowledge learning progress and regression rate is 0.6, and the classroom participation is 30 times (assuming that the number of speeches is used as the measurement standard). In this way, a teacher teaching supervision quality data report (Excel table) is generated to facilitate subsequent data analysis.

[0062] Preferably, a linear analysis visualization is performed on the relationship between each indicator score and the knowledge learning progress or classroom participation rate in the teacher teaching supervision quality data report to establish a linear visualization chart between each indicator score and the knowledge learning progress or classroom participation rate to generate corresponding classroom teaching supervision visualization results.

[0063] In an embodiment of the present invention, by using Python's matplotlib library and seaborn library, a linear analysis visualization is performed between each indicator score in the teacher teaching supervision quality data report and the knowledge learning progress and retreat rate or classroom participation. First, the scatter() function of the matplotlib library is used to draw a scatter plot, and the teaching method effectiveness score in the teacher teaching quality evaluation index is used as the horizontal coordinate, and the knowledge learning progress and retreat rate is used as the vertical coordinate. A scatter plot is drawn to observe the distribution trend between the two, and then the regplot() function of the seaborn library is used to add a linear regression fitting line to the scatter plot to intuitively display the linear relationship between the two. For example, through the scatter plot and the fitting line, it is found that the higher the teaching method effectiveness score, the higher the knowledge learning progress and retreat rate. In the same way, similar linear analysis visualization is performed on other indicators such as classroom interaction effect and knowledge learning progress and retreat rate, student learning achievement improvement and knowledge learning progress and retreat rate, and knowledge learning progress and retreat rate, as well as various indicators and classroom participation, to generate multiple linear visualization charts, which are sorted and summarized to generate corresponding classroom teaching supervision visualization results, providing intuitive data display and analysis basis for teaching supervision.

[0064] Furthermore, the present invention also provides a real-time dynamic learning data monitoring and visualization method, which is implemented based on the real-time dynamic learning data monitoring and visualization platform as described above, and the real-time dynamic learning data monitoring and visualization method includes: Through the combination of teaching visual data docking, learning and examination data docking, and teaching environment data docking and teaching basic status data warehouse, a corresponding teaching basic status collection model is constructed; the corresponding dynamic learning data is collected and monitored in real time and synchronously by using the docking corresponding to the teaching basic status collection model, and reported to the corresponding teaching basic status data warehouse; By conducting teaching modeling feature analysis on dynamic learning data in the teaching basic status data warehouse, a classroom teaching visual behavior feature set, a student learning behavior feature set, and a student environmental influencing factor set are obtained; based on the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environmental influencing factor set, personalized teaching analysis is conducted on the corresponding students to generate personalized teaching suggestions for classroom students; Based on the personalized teaching suggestions for students in class, the corresponding learning performance and classroom participation of students before and after the teaching recommendation are obtained, and the students' knowledge progress and regression analysis is performed on the corresponding learning performance before and after the teaching recommendation to obtain the corresponding knowledge learning progress and regression rate of students before and after the teaching recommendation; Based on dynamic learning data, the teaching supervision evaluation is visualized for students' knowledge learning progress and regression rates and classroom participation before and after teaching recommendations to generate corresponding classroom teaching supervision visualization results.

[0065] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0066] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A real-time dynamic learning data monitoring and visualization platform, characterized in that: Includes the following modules: The teaching status docking collection and reporting module is used to build a corresponding teaching basic status collection model through the combination of teaching affairs visual data docking, learning and examination affairs data docking, and teaching environment data docking with the teaching basic status data warehouse; the corresponding dynamic learning data is collected and monitored in real time and synchronously by using the docking corresponding to the teaching basic status collection model, and reported to the corresponding teaching basic status data warehouse; The teaching state modeling and analysis module is used to perform teaching modeling feature analysis on dynamic learning data in the teaching basic state data warehouse to obtain a classroom teaching visual behavior feature set, a student learning behavior feature set, and a student environmental influencing factor set; based on the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environmental influencing factor set, a personalized teaching analysis is performed on the corresponding students to generate personalized teaching suggestions for classroom students; The pre- and post-teaching recommendation analysis module is used to obtain the students' learning performance and class participation before and after the teaching recommendation based on the personalized teaching suggestions for the students in the classroom, and to analyze the students' knowledge progress and regression on their learning performance before and after the teaching recommendation to obtain the students' knowledge learning progress and regression rate before and after the teaching recommendation; The teaching supervision and evaluation visualization module is used to visualize the teaching supervision and evaluation of students' knowledge learning progress and regression rates and classroom participation before and after teaching recommendations based on dynamic learning data, so as to generate corresponding classroom teaching supervision visualization results.

2. The real-time dynamic learning data monitoring and visualization platform according to claim 1 is characterized in that: The teaching status docking collection and reporting module includes the following functions: By connecting the teaching affairs visual data, learning and examination affairs data, and teaching environment data with the teaching basic status data warehouse, a corresponding teaching basic status acquisition model is constructed; By configuring the corresponding teaching status synchronization collection time point, and based on the teaching status synchronization collection time point, using the corresponding docking of the teaching basic status collection model to monitor and collect the corresponding dynamic learning data in real time, including classroom dynamic visual data, learning dynamic behavior data and teaching dynamic environment data; Stream-process the corresponding dynamic learning data to generate a basic teaching state data stream; The basic teaching status data stream is transmitted and reported to the data sub-database in the basic teaching status data warehouse corresponding to the basic teaching status acquisition model.

3. The real-time dynamic learning data monitoring and visualization platform according to claim 2 is characterized in that: The method of using the teaching basic state acquisition model to collect the corresponding dynamic learning data based on the teaching state synchronous acquisition time point in real time synchronous monitoring includes: Based on the teaching status synchronous collection time point, the teaching affairs visual data corresponding to the basic teaching status collection model is used to connect the real-time monitoring collection of the corresponding classroom dynamic visual data, including the classroom panorama, teacher teaching activities and student classroom performance corresponding video stream data; Based on the teaching status synchronous collection time point, the learning and examination data corresponding to the teaching basic status collection model is connected to the real-time monitoring and collection of the corresponding learning dynamic behavior data, including the number of course visits, learning time, homework submission time, online test scores, students' corresponding body movements and sitting status in class; Based on the teaching status synchronous collection time point, the teaching environment data corresponding to the teaching basic status collection model is used to connect the real-time monitoring and collection of the corresponding teaching dynamic environment data, including the teacher's teaching content, students' questions and answers, classroom discussion sound corresponding voice text data and classroom learning environment data. The classroom learning environment data includes the corresponding temperature, humidity and light intensity of the classroom.

4. The real-time dynamic learning data monitoring and visualization platform according to claim 3 is characterized in that: The teaching status modeling and analysis module includes the following functions: By combining convolutional neural networks and recurrent neural networks in the teaching basic status data warehouse to build a corresponding classroom visual behavior analysis model, and obtaining the corresponding classroom teacher and student behavior annotation data, the classroom visual behavior analysis model is trained based on the classroom teacher and student behavior annotation data to generate a classroom behavior analysis model that can identify the corresponding teacher and student behaviors; Input the classroom dynamic visual data into the classroom behavior analysis model that can identify the corresponding teacher and student behaviors to perform classroom visual behavior feature analysis, so as to identify the corresponding teaching behavior features of teachers such as explanation, demonstration and questioning, and the corresponding classroom behavior features of students such as attentive listening, whispering, discussion and absent-mindedness, and obtain the classroom teaching visual behavior feature set; The spatiotemporal convolutional network is used to analyze the spatiotemporal variation characteristics of the course access times, learning time, homework submission time, and online test scores corresponding to the learning dynamic behavior data, so as to analyze the variation characteristics of the corresponding student behaviors in the time and space sequence, and the student concentration statistics are performed on the corresponding body movements and sitting postures of the students in the classroom in the learning dynamic behavior data to obtain the corresponding concentration of the students in the classroom. At the same time, the variation characteristics of the student behaviors in the time and space sequence and the corresponding concentration of the students in the classroom are combined to obtain the student learning behavior feature set; Based on the dynamic teaching environment data, the learning impact assessment of students' corresponding concentration in the classroom is carried out to determine the environmental factors that affect students' concentration during the teacher's classroom teaching process, so as to obtain the student environmental influence factor set; key mining and dimensionality reduction processing are carried out between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influence factor set, so as to generate the dimensionality reduction feature set of the basic teaching state; Through the dimensionality reduction feature set analysis of the basic teaching status, the corresponding classroom learning evaluation of the students is obtained, including the corresponding learning history, interest preferences and knowledge mastery of the students. Personalized teaching analysis is performed based on the corresponding classroom learning evaluation of the students, so as to recommend corresponding learning resources and learning paths for the students' weak knowledge points and generate personalized teaching suggestions for students in the classroom.

5. The real-time dynamic learning data monitoring and visualization platform according to claim 4 is characterized in that: The student concentration statistics of the corresponding body movements and sitting postures of the students in the classroom in the learning dynamic behavior data include: Performing time-series synchronization processing on the corresponding body movements and sitting postures of students in the classroom in the learning dynamic behavior data to generate corresponding student body movement sequences and student sitting posture sequences within the same time-series range; The corresponding student head rotation angle and student arm swing angle are obtained through the student body movement sequence, and the center of gravity shift analysis is performed according to the student head rotation angle and the student arm swing angle to obtain the degree of student center of gravity shift; The duration of the students' gaze at the blackboard is obtained through the student's body movement sequence, and the number of twists of the sitting body is obtained through the student's sitting state sequence; The sitting posture stability evaluation is calculated based on the duration of the students staring at the blackboard and the number of twists of the body in the sitting position, so as to obtain the degree of students' sitting posture stability; The students' concentration is counted based on the degree of their center of gravity shift and the stability of their sitting posture to obtain the corresponding concentration of the students in class.

6. The real-time dynamic learning data monitoring and visualization platform according to claim 4, characterized in that: The learning impact assessment and judgment of the student's corresponding concentration in class based on the teaching dynamic environment data includes: Based on the various environmental factors in the teaching dynamic environment data, a linear correlation impact assessment is performed on the corresponding concentration of students in the classroom to obtain the linear correlation influence coefficient between each teaching environment factor and concentration; Based on the linear correlation influence coefficient between each teaching environment factor and the concentration, the learning impact judgment and screening of each environmental factor in the teaching dynamic environment data is carried out. If the linear correlation influence coefficient between the corresponding teaching environment factor and the concentration is positively correlated, then the corresponding teaching environment factor is the environmental factor that affects the student's concentration; if the linear correlation influence coefficient between the corresponding teaching environment factor and the concentration is negatively correlated, then the corresponding teaching environment factor is not the environmental factor that affects the student's concentration. In this way, the environmental factors that affect the student's concentration during the teacher's classroom teaching process are judged and screened to obtain the student environment influencing factor set.

7. The real-time dynamic learning data monitoring and visualization platform according to claim 4, characterized in that: The key mining and dimensionality reduction processing between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influencing factor set includes: The correlation measurement calculation is performed between the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influencing factor set, so as to quantitatively calculate the correlation coefficient between the two sub-features and form a corresponding correlation matrix to generate the basic state feature correlation matrix of teaching; Based on the correlation matrix of basic teaching status characteristics, the corresponding sub-features in the classroom teaching visual behavior feature set, the student learning behavior feature set and the student environmental influencing factor set are screened for behavioral feature key mining. If the correlation coefficient between the two sub-features in the correlation matrix of basic teaching status characteristics is greater than 0.75, it is considered that the two sub-features are all key features; if the correlation coefficient between the two sub-features in the correlation matrix of basic teaching status characteristics is equal to 0.75, it is considered that there is a key feature in the two sub-features; if the correlation coefficient between the two sub-features in the correlation matrix of basic teaching status characteristics is less than 0.75, it is considered that the two sub-features are not key features; select the corresponding key features according to the above judgment rules to generate the key feature set of basic teaching status; The principal component dimensionality reduction processing is performed on the corresponding sub-features in the key feature set of the basic teaching state to generate a dimensionality reduction feature set of the basic teaching state.

8. The real-time dynamic learning data monitoring and visualization platform according to claim 1, characterized in that: The teaching recommendation before and after analysis module includes the following functions: Based on the personalized teaching suggestions for students in class, the corresponding classroom teaching recommendations are implemented, and the corresponding learning performance and classroom participation of students before and after the teaching recommendations are obtained, where the learning performance includes the learning performance corresponding to each subject knowledge; Perform learning progress and regression space analysis on each subject knowledge within the students' corresponding learning scores before and after the teaching recommendation, so as to quantify the difference in learning scores of each subject knowledge before and after the teaching recommendation, and obtain the learning progress and regression space corresponding to each subject knowledge; The learning progress and regression range of each subject knowledge in the students' corresponding learning scores before and after the teaching recommendation is calculated, so as to quantify the ratio of the corresponding learning scores of each subject knowledge before and after the teaching recommendation, and obtain the learning progress and regression range corresponding to each subject knowledge; The students' knowledge progress and regression are summarized and calculated based on the learning progress and regression space and the learning progress and regression amplitude corresponding to each subject knowledge, so as to obtain the students' corresponding knowledge learning progress and regression rate before and after the teaching recommendation.

9. The real-time dynamic learning data monitoring and visualization platform according to claim 3, characterized in that: The teaching supervision and evaluation visualization module includes the following functions: According to the classroom dynamic visual data, learning dynamic behavior data and teaching dynamic environment data corresponding to the dynamic learning data, a teaching supervision evaluation index system is constructed to generate corresponding teacher teaching quality evaluation indicators, student learning effect evaluation indicators and teaching environment evaluation indicators. The teacher teaching quality evaluation indicators include the effectiveness of teaching methods, classroom interaction effects and the improvement of student learning outcomes. The student learning effect evaluation indicators include academic performance and learning interest cultivation. The teaching environment evaluation indicators include classroom learning impact, classroom facility applicability and environmental learning atmosphere. Based on the scores of the teacher teaching quality evaluation indicators, student learning effect evaluation indicators and teaching environment evaluation indicators, a teaching supervision data report is established for the students' knowledge learning progress and regression rate and classroom participation before and after the teaching recommendation, so as to generate a teacher teaching supervision quality data report; Perform linear analysis and visualization on the relationship between each indicator score and the knowledge learning progress or classroom participation rate in the teacher teaching supervision quality data report, so as to establish a linear visualization chart between each indicator score and the knowledge learning progress or classroom participation rate, so as to generate the corresponding classroom teaching supervision visualization results.

10. A real-time dynamic learning data monitoring and visualization method, characterized in that: The method is implemented based on the real-time dynamic learning data monitoring and visualization platform according to claim 1, and the real-time dynamic learning data monitoring and visualization method comprises: Through the combination of teaching visual data docking, learning and examination data docking, and teaching environment data docking and teaching basic status data warehouse, a corresponding teaching basic status collection model is constructed; the corresponding dynamic learning data is collected and monitored in real time and synchronously by using the docking corresponding to the teaching basic status collection model, and reported to the corresponding teaching basic status data warehouse; By conducting teaching modeling feature analysis on dynamic learning data in the teaching basic status data warehouse, a classroom teaching visual behavior feature set, a student learning behavior feature set, and a student environmental influencing factor set are obtained; based on the classroom teaching visual behavior feature set, the student learning behavior feature set, and the student environmental influencing factor set, personalized teaching analysis is conducted on the corresponding students to generate personalized teaching suggestions for classroom students; Based on the personalized teaching suggestions for students in class, the corresponding learning performance and classroom participation of students before and after the teaching recommendation are obtained, and the students' knowledge progress and regression analysis is performed on the corresponding learning performance before and after the teaching recommendation to obtain the corresponding knowledge learning progress and regression rate of students before and after the teaching recommendation; Based on dynamic learning data, the teaching supervision evaluation is visualized for students' knowledge learning progress and regression rates and classroom participation before and after teaching recommendations to generate corresponding classroom teaching supervision visualization results.

Images