A two-way monitoring system for student entertainment learning

Abstract

The application discloses a kind of for student entertainment learning two-way supervision system, belong to educational supervision field, the system includes: label mark module, monitoring management module and early warning intervention module;Label mark module is split into sub-period and assigns specific matter with monitoring period, constructs characteristic vector by collecting multidimensional characteristic data, divides learning and entertainment sub-label, determines the sub-label corresponding to each unit time by characteristic matching value;Monitoring management module calculates the correlation value of sub-label and matter and is converted into fitting value, constructs change broken line and compares with standard broken line, divides fitting section and calculates total evaluation value, to judge whether trigger early warning;Early warning intervention module filters low fitting section sub-label to form inconsistent label set when early warning, constructs diagnostic radar chart and matches inducement and takes targeted intervention measure;The system realizes the fine identification of student state, dynamic quantitative evaluation and accurate intervention, effectively improves the scientific nature and individualized level of supervision.

Description

Technical Field

[0001] This invention belongs to the field of educational supervision technology, specifically relating to a two-way supervision system for students' entertainment and learning. Background Technology

[0002] Traditional education supervision often focuses on a single dimension, lacking a comprehensive consideration of students' learning and leisure activities, making it difficult to promptly identify imbalances between learning and leisure. Specifically: 1. It is difficult to accurately identify the specific state type of learning and entertainment; 2. Lack of dynamic quantitative assessment of the match between status and activity makes it impossible to determine whether the status meets expectations; 3. After issuing an early warning, it is difficult to accurately identify the cause and take personalized measures, resulting in poor intervention effects and delayed and ineffective intervention measures. To address this, we propose a two-way monitoring system for students' entertainment and learning. Summary of the Invention

[0003] The purpose of this invention is to provide a two-way monitoring system for student entertainment and learning, in order to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a two-way monitoring system for student entertainment and learning, comprising: a tag marking module, a monitoring and management module, and an early warning and intervention module; Tag marking module: Divide the entertainment and learning monitoring period into multiple monitoring sub-periods, assign specific tasks to each period, collect multi-dimensional feature data within each monitoring sub-period, construct feature vectors for each dimension of features, divide learning and entertainment sub-tags, determine the associated features of each sub-tag, calculate feature matching values, and determine the sub-tags corresponding to each unit of time accordingly. Monitoring and Management Module: For each monitoring sub-period, calculate the correlation value between the sub-tag and the specific item and convert it into a matching value. Construct a matching value change line and compare it with the standard line. Divide the matching segments into standard, low and high matching segments, calculate the total matching score, and determine whether to trigger an alert based on this score. Early warning and intervention module: When a monitoring early warning is triggered, low-matching joke tags are filtered to form a set of mismatched tags, their frequency is counted and a diagnostic radar chart is constructed; the diagnostic radar chart is used to match the diagnostic causes and intervention measures are taken accordingly.

[0005] Preferably, the specific process for constructing feature vectors for each dimension is as follows: Based on students' learning tasks and daily routines, the entertainment and learning monitoring period is divided into multiple monitoring sub-periods; and specific tasks are assigned to each monitoring sub-period. For each monitoring sub-period, collect multi-dimensional characteristic data of students; For each dimension of feature data, integrate them into an independent data unit; Set learning status tags and entertainment status tags, and each of the learning status tags contains several sub-tags; For each feature dimension, key statistical indicators for that feature dimension are extracted to form a feature vector for that feature dimension.

[0006] Preferably, the specific process for determining the association features of each sub-label is as follows: Calculate the absolute value of the correlation coefficient between the feature vector corresponding to each dimension and each sub-label. If the absolute value of the correlation coefficient between the feature vector and a certain state label is greater than the corresponding threshold, then mark the dimension feature corresponding to the feature vector as the associated feature of the corresponding sub-label. For each sub-label, based on the correlation coefficient between each associated feature and the corresponding sub-label, the relational features are divided into positively correlated features and negatively correlated features.

[0007] Preferably, the specific process for determining the sub-labels corresponding to each unit of time is as follows: For each associated feature, the frequency, duration, and intensity of occurrence per unit time are calculated, and after normalization, a comprehensive analysis is performed to obtain the feature matching value. A preset sub-label feature matching value range is defined. If the feature matching value corresponding to the current unit time of a sub-label is within the corresponding sub-label feature matching value range, then that unit time is marked as the corresponding sub-label time period.

[0008] Preferably, the specific process for constructing the broken line of fit value changes is as follows: For each monitoring sub-period, calculate the correlation value between each sub-label and the specific items in the monitoring sub-period; Set several relevance value ranges, and pre-determine a matching value for each relevance value range; match the relevance value corresponding to each sub-tag with all relevance value ranges to determine the corresponding matching value; A two-dimensional rectangular coordinate system is constructed with the alignment value as the vertical axis and time as the horizontal axis to obtain the sub-labels corresponding to each unit of time and the alignment value between the sub-labels and specific items in the current monitoring sub-period in real time. The sub-labels corresponding to each unit time are labeled in a Cartesian coordinate system according to their corresponding matching values, resulting in several sets of matching value-time data pairs. Based on the chronological order, data pairs corresponding to adjacent time units are connected sequentially by a line to obtain a line graph showing the change in the fit value.

[0009] The preferred process for dividing the segments into standard, low, and high matching segments is as follows: The line segment that coincides with the line segment of the fit value change and the line segment of the standard fit value change is marked as the standard fit segment; Line segments whose fit value change line is lower than the standard fit value change line are marked as low fit segments; The segments of the line graph showing the change in fit value that are higher than the standard line graph showing the change in fit value are marked as high fit segments.

[0010] Preferably, the specific process for calculating the overall matching score and determining whether to trigger an alert is as follows: Calculate the proportions of the standard fit segment, low fit segment, and high fit segment in the line graph of fit value change up to the current unit of time, and obtain the standard fit ratio BZ, low fit ratio DZ, and high fit ratio GZ. Calculate the average deviation between the fit value and the standard fit value for the low fit segment and the high fit segment respectively, and obtain the low fit deviation DP and the high fit deviation GP. After normalizing the standard fit percentage (BZ), low fit percentage (DZ), high fit percentage (GZ), low fit deviation (DP), and high fit deviation (GP) corresponding to the current unit time, the total fit score (QPZ) is obtained using the formula: QPZ=BZ×a1+GZ×a2+GP×a3-DZ×a4-DP×a5; where a1, a2, a3, a4, and a5 are preset weighting coefficients. The overall matching score for the current unit of time is compared with the corresponding basic threshold. If the overall matching score is less than the corresponding basic threshold, a monitoring warning is triggered.

[0011] Preferably, the specific process for constructing a diagnostic radar chart is as follows: For each specific item, when a monitoring alert is triggered, the sub-tags in the low-fit segment are filtered out; a set of non-matching tags is formed, and the frequency of each sub-tag in the set of non-matching tags is counted. Based on all potential sub-label types covered by this specific matter, a multi-axis radar chart is constructed; each axis in the radar chart corresponds to a sub-label type, and on each axis of the radar chart, a series of equally spaced scale markings are set according to the frequency range of possible occurrence of the sub-label. Each sub-label and its corresponding quantity in the mismatch label set are marked on the radar diagnostic chart; and the scales marked on each coordinate axis are connected in a preset order to obtain a closed shape, which is called the diagnostic radar chart.

[0012] Preferably, the specific process for matching and diagnosing the underlying cause is as follows: Construct a diagnostic image library, which contains several simulated radar images; each simulated radar image is pre-defined to correspond to a specific diagnostic trigger. The diagnostic radar image corresponding to the triggered monitoring and early warning is substituted into the diagnostic image library and its similarity is calculated with each simulated radar image. The diagnostic cause corresponding to the simulated radar image with the largest similarity value that is greater than the preset threshold is output, and corresponding intervention measures are taken based on the diagnostic cause.

[0013] Compared with the prior art, the beneficial effects of the present invention are: (1) This two-way monitoring system for student entertainment and learning provides rich data support for comprehensive analysis of student status by collecting multi-dimensional characteristic data of students. Based on students' learning tasks and life patterns, the monitoring period for entertainment and learning is divided into multiple monitoring sub-periods, and specific items are assigned to each period. This makes the monitoring of student status more targeted and systematic. For each dimension of characteristic data, key statistical indicators are extracted to form a feature vector. By calculating the feature matching value, the sub-labels corresponding to each unit of time are determined, realizing the quantitative assessment and automatic judgment of student status. This not only improves the objectivity and accuracy of the judgment, but also enables rapid response to changes in student status. It breaks through the limitations of traditional monitoring that is rough in depicting student status and difficult to distinguish subdivided statuses, and provides accurate basis for subsequent monitoring management and early warning intervention.

[0014] (2) This two-way monitoring system for student entertainment and learning calculates the correlation value between sub-labels and specific items and converts it into a matching value. It constructs a matching value change line and compares it with the standard line to accurately divide the standard, low and high matching segments. It calculates the comprehensive total score by combining the proportion of each segment and the deviation amount, and judges in real time whether the student's status meets the expectations. It solves the problems of traditional monitoring lacking scientific quantitative assessment, relying on subjective judgment and being lagging behind, and makes status assessment more objective, dynamic and accurate.

[0015] (3) This two-way monitoring system for student entertainment and learning forms a precise diagnosis and personalized closed-loop management from early warning to intervention. When an early warning is triggered, low-matching joke tags are selected to construct a diagnostic radar chart. By matching the simulated radar chart in the diagnostic chart library, specific causes such as environmental interference and lack of learning interest are determined, and then targeted measures are taken to change the general and lagging status of traditional intervention measures. It realizes precise and personalized management of the whole process from abnormal status identification to problem solving, effectively balances students' learning and entertainment, and improves the effectiveness of supervision. Attached Figure Description

[0016] Figure 1 This is a flowchart of the present invention; Figure 2 This is a diagnostic radar chart for the present invention. Detailed Implementation

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

[0018] Example 1 Please see Figure 1 - Figure 2 This invention provides a two-way monitoring system for student entertainment and learning, comprising: a tag marking module, a monitoring and management module, and an early warning and intervention module; The tagging module divides the entertainment and learning monitoring period into multiple monitoring sub-periods, assigns specific tasks to each period, collects multi-dimensional feature data within each monitoring sub-period, constructs feature vectors for each dimension of features, divides the learning and entertainment sub-tags, determines the associated features of each sub-tag, calculates feature matching values, and determines the sub-tag corresponding to each unit of time accordingly. The specific process is as follows: Based on students' learning tasks and daily routines, the entertainment and learning monitoring period is divided into multiple monitoring sub-periods; the monitoring sub-periods include: course learning period, knowledge expansion period, leisure and relaxation period, and self-study period; Assign specific tasks to each monitoring sub-period, specifically: Course learning sessions include: mathematical logic training, Chinese reading comprehension, foreign language listening practice, scientific experimental inquiry, and historical event discussion. Knowledge expansion period: Subject knowledge competitions, popular science lectures, literary appreciation, cultural and artistic experiences, and other extension activities; Leisure and relaxation time: entertainment options such as brain games, fun reading, music appreciation, sports activities, and social interaction; Self-study period: Learning activities include completing assignments, conducting project research, self-review, and previewing lessons; For each monitoring sub-period, multi-dimensional characteristic data of students are collected; the multi-dimensional characteristic data of students includes: physical characteristic data, voice characteristic data, environmental characteristic data, and interaction characteristic data; For each dimension of feature data, it is integrated into an independent data unit; the data unit includes key information such as timestamp, feature type, and quantization value. Physical characteristic data include: students' facial micro-expressions, eye movements, head rotation angle, and the amplitude and frequency of limb movements; Speech feature data includes: purity of student voice timbre, amplitude of intonation fluctuation, rate of speech rate change, duration and frequency of pauses; Environmental characteristic data include: light color temperature, ambient temperature fluctuation, and noise spectrum distribution; Interaction feature data includes: the interactive behaviors between students and learning devices, entertainment props, and social objects; Set learning status tags and entertainment status tags, and each of the learning status tags contains several sub-tags; The sub-tags corresponding to the learning status tags include: Deeply focused learning, difficulty in absorbing knowledge, active creative thinking, and knowledge consolidation and review. Sub-tags corresponding to the entertainment status tag include: immersive relaxation, socially active and joyful, entertainment challenge and focused, and emotional relaxation and recovery. The process involves analyzing the multi-dimensional feature data collected within the current unit of time to determine the sub-label corresponding to the current unit of time. Perform data cleaning and data normalization preprocessing on the collected multi-dimensional feature data; For each feature dimension, key statistical indicators are extracted to form the feature vector of the corresponding feature dimension; key statistical indicators include: mean, standard deviation, maximum, minimum and other statistical measures; Calculate the absolute value of the correlation coefficient between the feature vector corresponding to each dimension feature and each sub-label. Preset the correlation coefficient threshold. If the absolute value of the correlation coefficient between the feature vector and a certain state label is greater than the corresponding threshold, then mark the dimension feature corresponding to the feature vector as the associated feature of the corresponding sub-label. For each sub-label, based on the correlation coefficient between each associated feature and the corresponding sub-label, the relational features are divided into positively correlated features and negatively correlated features; For each associated feature, calculate its frequency, duration, and intensity per unit time to obtain the feature occurrence frequency (TP), feature duration (TS), and feature intensity value (TQ). After normalization, use the formula: The feature matching value PT is obtained; Where i is the label of the positively correlated feature, n is the total number of positively correlated features; j is the label of the negatively correlated feature, and m is the total number of negatively correlated features; , , These are the preset weighting coefficients for the frequency of feature occurrence, the duration of feature occurrence, and the intensity value of feature in the i-th positively correlated feature; , , These are the preset weighting coefficients for the frequency of feature occurrence, the duration of feature occurrence, and the intensity value of feature in the j-th negatively correlated feature; A preset sub-label feature matching value range is defined. If the feature matching value corresponding to the current unit time of a sub-label is within the corresponding sub-label feature matching value range, then that unit time is marked as the corresponding sub-label time period.

[0019] It should be noted that by collecting multi-dimensional characteristic data of students, including physical, voice, environmental and interactive aspects, rich data support is provided for a comprehensive analysis of students' status; this data can reflect students' learning and entertainment status in different monitoring sub-periods, which helps to achieve accurate monitoring of students' behavior and emotions; For example, facial micro-expressions and eye movements in body feature data can reflect students' attention and emotional state; intonation and speech rate in voice feature data can reflect students' participation and level of mental activity; environmental feature data helps to understand the impact of learning and entertainment environments on students' state; and interaction feature data can reflect students' interaction with their surroundings. Based on students' learning tasks and daily routines, the monitoring period for entertainment and learning is divided into multiple monitoring sub-periods, and specific tasks are assigned to each period. This division method makes the monitoring of students' status more targeted and systematic, and can better adapt to the rhythm and characteristics of students' daily activities, providing a clear time framework for subsequent analysis and intervention. For each dimension of feature data, key statistical indicators are extracted to form feature vectors, which realizes a concise representation and effective integration of complex data. Feature vectors can highlight the main features and patterns of the data, which facilitates subsequent calculations and analysis, and improves the efficiency and accuracy of data processing. Setting learning status labels and entertainment status labels, and further subdividing them into several sub-labels, while determining the correlation characteristics of each sub-label; this process helps to classify students' status in a more refined way, enabling the system to more accurately identify and understand students' behavior and internal state in different situations, and providing a basis for personalized intervention measures. By calculating feature matching values ​​and pre-setting sub-label feature matching value ranges to determine the sub-labels corresponding to each unit of time, quantitative assessment and automatic judgment of student status are achieved. This method not only improves the objectivity and accuracy of judgment, but also enables rapid response to changes in student status, providing accurate basis for subsequent monitoring management and early warning intervention.

[0020] The monitoring and management module calculates the correlation value between sub-tags and specific items for each monitoring sub-period and converts it into a matching value. It then constructs a matching value change line and compares it with a standard line, dividing the system into standard, low, and high matching segments. Finally, it calculates the overall matching score and determines whether to trigger an alert. The specific process is as follows: For each monitoring sub-period, calculate the correlation value between each sub-tag and the specific items of the monitoring sub-period, set several correlation value intervals, and pre-set that each correlation value interval corresponds to a matching value. Match the relevance value of each sub-tag with all relevance value ranges to determine the corresponding fit value; The correlation value can be positive or negative. If the correlation value between the sub-label and the specific item in the monitoring sub-period is positive and the larger the correlation value, the larger the corresponding correlation value. Conversely, if the correlation value between the sub-label and the specific item in the monitoring sub-period is negative and the absolute value of the correlation is large, the corresponding correlation value is small. A two-dimensional rectangular coordinate system is constructed with the alignment value as the vertical axis and time as the horizontal axis to obtain the sub-labels corresponding to each unit of time and the alignment value between the sub-labels and specific items in the current monitoring sub-period in real time. The sub-labels corresponding to each unit time are labeled in a Cartesian coordinate system according to their corresponding matching values, resulting in several sets of matching value-time data pairs. Based on the chronological order, data pairs corresponding to consecutive time units are connected sequentially using a line graph to obtain a line graph showing the change in the matching value; at the same time, a standard line graph showing the change in the matching value for the current monitoring sub-time period is preset. The line segments that coincide with the standard fit value change line are marked as standard fit segments; the line segments whose fit value change line is lower than the standard fit value change line are marked as low fit segments; and the line segments whose fit value change line is higher than the standard fit value change line are marked as high fit segments. Calculate the proportions of the standard fit segment, low fit segment, and high fit segment in the line graph of fit value change up to the current unit of time, and obtain the standard fit ratio BZ, low fit ratio DZ, and high fit ratio GZ. Calculate the average deviation between the fit value and the standard fit value for the low fit segment and the high fit segment respectively, and obtain the low fit deviation DP and the high fit deviation GP. After normalizing the standard fit percentage (BZ), low fit percentage (DZ), high fit percentage (GZ), low fit deviation (DP), and high fit deviation (GP) corresponding to the current unit time, the total fit score (QPZ) is obtained using the formula: QPZ=BZ×a1+GZ×a2+GP×a3-DZ×a4-DP×a5; where a1, a2, a3, a4, and a5 are preset weighting coefficients. A preset baseline threshold for the overall fit score is established. The overall fit score corresponding to the current unit of time is compared with the corresponding baseline threshold. If the overall fit score is less than the corresponding baseline threshold, a monitoring alert is triggered.

[0021] It should be noted that by constructing a line graph showing the changes in the fit value, the system can display the trend of students' status changes in each monitoring sub-period in real time and intuitively; this allows teachers and parents to understand at any time whether students' learning and entertainment status meets expectations and to identify problems in a timely manner. When the overall score falls below a preset threshold, the system triggers a monitoring alert. This alert mechanism can promptly remind relevant personnel that a student may be in a poor state and requires intervention to prevent the problem from worsening. Converting the relevance values ​​of sub-labels to specific items into fit values ​​makes the assessment of student status more quantitative and intuitive; positive and negative fit values ​​can clearly reflect the degree of matching between student status and preset activity types, helping teachers and parents to better understand students' behavioral performance. By comparing the line graph of the fit value change with the standard line graph, the system can clearly divide the standard fit segment, the low fit segment, and the high fit segment. This comparative analysis helps to accurately locate the specific time period and degree to which the student's state deviates from the normal range, providing detailed basis for further diagnosis and intervention. Taking into account factors such as the proportion of students meeting the standard criteria, the proportion of students with low or high standards, and the deviations from the standard and high standards, a comprehensive fit score is calculated. This comprehensive evaluation index can fully reflect the overall status of students in the current monitoring sub-period, providing a scientific basis for the development of personalized intervention measures.

[0022] Early warning and intervention module: When a monitoring early warning is triggered, low-relevance joke tags are filtered to form a mismatch tag set, their frequency is counted, and a diagnostic radar chart is constructed; based on the diagnostic radar chart, diagnostic causes are matched, and intervention measures are taken accordingly. The specific process is as follows: For each specific item, when a monitoring alert is triggered, the sub-tags in the low-fit segment are filtered out; a set of non-matching tags is formed, and the frequency of each sub-tag in the set of non-matching tags is counted. Based on all potential sub-label types covered by this specific matter, a multi-axis radar chart is constructed; each axis in the radar chart corresponds to a sub-label type, and on each axis of the radar chart, a series of equally spaced scale markings are set according to the frequency range of possible occurrence of the sub-label. Each sub-label and its corresponding quantity in the mismatch label set is labeled on the radar diagnostic chart; and the scales on each coordinate axis are connected sequentially according to a preset order to obtain a closed graph, which is denoted as the diagnostic radar chart; for example... Figure 2 ; A diagnostic image library is constructed, which contains several simulated radar images. Each simulated radar image is pre-defined to correspond to a specific diagnostic trigger. Diagnostic triggers include: environmental interference factors, lack of interest in learning, insufficient rest time, etc. The diagnostic radar chart that triggers the monitoring and early warning is substituted into the diagnostic chart library and its similarity is calculated with each simulated radar chart. The diagnostic cause corresponding to the simulated radar chart with the largest similarity value that is greater than a preset threshold is output. Based on the diagnostic cause, corresponding intervention measures are taken to help students better balance entertainment and learning activities and improve the overall supervision effect.

[0023] It should be noted that when a monitoring alert is triggered, the system quickly filters low-matching joke tags to form a mismatch tag set and counts their frequency; this process can quickly identify the specific manifestations of a student's current state not matching the preset activity type, providing a clear direction for timely intervention; By constructing a diagnostic radar chart, the sub-labels in the non-conforming label set and their frequencies can be displayed intuitively. The visualization effect of the diagnostic radar chart helps teachers and parents quickly understand the deviation of students' status, and improves the pertinence and efficiency of intervention. By using simulated radar charts in the diagnostic library, the system can accurately identify the diagnostic triggers that lead to low fit. This data- and pre-set model-based diagnostic approach avoids errors from subjective judgment and improves the accuracy and reliability of the diagnosis. By taking corresponding intervention measures based on different diagnostic causes, personalized support for students can be achieved. For example, if the problem is caused by environmental interference factors, the learning environment can be adjusted; if the problem is a lack of interest in learning, the learning content or form can be enriched. This targeted approach helps to better meet the individual needs of students and improve the effectiveness of intervention. Through the regulatory role of the early warning and intervention module, the system helps students find a balance between entertainment and learning. When students are in a low-fit state during study time, the system adjusts the study schedule or methods in a timely manner. When students lack a positive state during entertainment time, the system can increase beneficial entertainment activities. This balance helps students achieve full development in both aspects and promotes the improvement of their comprehensive qualities. As the closed-loop terminal of the monitoring system, the early warning and intervention module transforms monitoring results into specific intervention actions. It not only completes real-time monitoring and early warning of students' status, but also further realizes the diagnosis and resolution of problems, making the entire monitoring system an organic whole and improving the overall effectiveness of the system. The tagging module, monitoring and management module, and early warning and intervention module work together to achieve real-time monitoring, early warning, and intervention of students' entertainment and learning status. This complete closed-loop supervision model ensures that students' learning and entertainment activities are always under effective management, and any deviation from the normal state can be corrected in a timely manner.

[0024] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A two-way monitoring system for student entertainment and learning, comprising: The tag marking module, monitoring and management module, and early warning and intervention module are characterized by: Tag marking module: Divide the entertainment and learning monitoring period into multiple monitoring sub-periods, assign specific tasks to each period, collect multi-dimensional feature data within each monitoring sub-period, construct feature vectors for each dimension of features, divide learning and entertainment sub-tags, determine the associated features of each sub-tag, calculate feature matching values, and determine the sub-tags corresponding to each unit of time accordingly. Monitoring and Management Module: For each monitoring sub-period, calculate the correlation value between the sub-tag and the specific item and convert it into a matching value. Construct a matching value change line and compare it with the standard line. Divide the matching segments into standard, low and high matching segments, calculate the total matching score, and determine whether to trigger an alert based on this score. Early warning and intervention module: When a monitoring early warning is triggered, low-matching joke tags are filtered to form a set of mismatched tags, their frequency is counted and a diagnostic radar chart is constructed; the diagnostic radar chart is used to match the diagnostic causes and intervention measures are taken accordingly.

2. The two-way monitoring system for student entertainment and learning according to claim 1, characterized in that: The specific process of constructing feature vectors for each dimension is as follows: Based on students' learning tasks and daily routines, the entertainment and learning monitoring period is divided into multiple monitoring sub-periods; and specific tasks are assigned to each monitoring sub-period. For each monitoring sub-period, collect multi-dimensional characteristic data of students; For each dimension of feature data, integrate them into an independent data unit; Set learning status tags and entertainment status tags, and each of the learning status tags contains several sub-tags; For each feature dimension, key statistical indicators for that feature dimension are extracted to form a feature vector for that feature dimension.

3. The two-way monitoring system for student entertainment and learning according to claim 2, characterized in that: The specific process for determining the association features of each sub-label is as follows: Calculate the absolute value of the correlation coefficient between the feature vector corresponding to each dimension and each sub-label. If the absolute value of the correlation coefficient between the feature vector and a certain state label is greater than the corresponding threshold, then mark the dimension feature corresponding to the feature vector as the associated feature of the corresponding sub-label. For each sub-label, based on the correlation coefficient between each associated feature and the corresponding sub-label, the relational features are divided into positively correlated features and negatively correlated features.

4. The two-way monitoring system for student entertainment and learning according to claim 3, characterized in that: The specific process for determining the sub-labels corresponding to each unit of time is as follows: For each associated feature, the frequency, duration, and intensity of occurrence per unit time are calculated, and after normalization, a comprehensive analysis is performed to obtain the feature matching value. A preset sub-label feature matching value range is defined. If the feature matching value corresponding to the current unit time of a sub-label is within the corresponding sub-label feature matching value range, then that unit time is marked as the corresponding sub-label time period.

5. A two-way monitoring system for student entertainment and learning according to claim 4, characterized in that: The specific process of constructing the line graph of the fit value change is as follows: For each monitoring sub-period, calculate the correlation value between each sub-label and the specific items in the monitoring sub-period; Set several relevance value ranges, and pre-determine a matching value for each relevance value range; match the relevance value corresponding to each sub-tag with all relevance value ranges to determine the corresponding matching value; A two-dimensional rectangular coordinate system is constructed with the alignment value as the vertical axis and time as the horizontal axis to obtain the sub-labels corresponding to each unit of time and the alignment value between the sub-labels and specific items in the current monitoring sub-period in real time. The sub-labels corresponding to each unit of time are labeled in a rectangular coordinate system according to their corresponding matching values, resulting in several sets of matching value-time data pairs. Based on the chronological order, data pairs corresponding to adjacent time units are connected sequentially by a line to obtain a line graph showing the change in the fit value.

6. A two-way monitoring system for student entertainment and learning according to claim 5, characterized in that: The specific process of dividing the standard, low, and high matching segments is as follows: The line segment that coincides with the line segment of the fit value change and the line segment of the standard fit value change is marked as the standard fit segment; Line segments whose fit value change line is lower than the standard fit value change line are marked as low fit segments; The segments of the line graph showing the change in fit value that are higher than the standard line graph showing the change in fit value are marked as high fit segments.

7. A two-way monitoring system for student entertainment and learning according to claim 6, characterized in that: The specific process for calculating the overall matching score and determining whether to trigger an alert is as follows: Calculate the proportions of the standard fit segment, low fit segment, and high fit segment in the line graph of fit value change up to the current unit of time, and obtain the standard fit ratio BZ, low fit ratio DZ, and high fit ratio GZ. Calculate the average deviation between the fit value and the standard fit value for the low fit segment and the high fit segment respectively, and obtain the low fit deviation DP and the high fit deviation GP. After normalizing the standard fit percentage (BZ), low fit percentage (DZ), high fit percentage (GZ), low fit deviation (DP), and high fit deviation (GP) corresponding to the current unit time, the total fit score (QPZ) is obtained using the formula: QPZ=BZ×a1+GZ×a2+GP×a3-DZ×a4-DP×a5; where a1, a2, a3, a4, and a5 are preset weighting coefficients. The overall matching score for the current unit of time is compared with the corresponding basic threshold. If the overall matching score is less than the corresponding basic threshold, a monitoring warning is triggered.

8. A two-way monitoring system for student entertainment and learning according to claim 7, characterized in that: The specific process of constructing a diagnostic radar chart is as follows: For each specific item, when a monitoring alert is triggered, the sub-tags in the low-fit segment are filtered out; a set of non-matching tags is formed, and the frequency of each sub-tag in the set of non-matching tags is counted. Construct a multi-axis radar chart based on all potential sub-label types covered by this specific item; Each axis in the radar chart corresponds to a sub-label type, and on each axis of the radar chart, a series of equally spaced scale markings are set according to the frequency range of possible occurrence of the sub-label. Each sub-label and its corresponding quantity in the mismatch label set are marked on the radar diagnostic chart; and the scales marked on each coordinate axis are connected in a preset order to obtain a closed shape, which is called the diagnostic radar chart.

9. A two-way monitoring system for student entertainment and learning according to claim 8, characterized in that: The specific process of matching and diagnosing the underlying cause is as follows: Construct a diagnostic image library, which contains several simulated radar images; each simulated radar image is pre-defined to correspond to a specific diagnostic trigger. The diagnostic radar image corresponding to the triggered monitoring and early warning is substituted into the diagnostic image library and its similarity is calculated with each simulated radar image. The diagnostic cause corresponding to the simulated radar image with the largest similarity value that is greater than the preset threshold is output, and corresponding intervention measures are taken based on the diagnostic cause.

Images