Intelligent education system

Abstract

The embodiment of the invention provides a wisdom education system, and belongs to the technical field of wisdom education. The system comprises an exception analysis module which analyzes first behavior data to obtain a first exceptional behavior and an exceptional position of the first exceptional behavior; the data obtaining module obtains adjacent behavior data from the first behavior data according to the abnormal position; the behavior prediction module performs behavior reconstruction according to the adjacent behavior data to obtain a target prediction behavior; the exception screening module screens out a second exception behavior according to the target prediction behavior and the first exception behavior; the exception classification module performs classification according to the second exception behavior to obtain an exception type; the exception determination module determines a target exception behavior according to the exception type, and obtains an associated teaching text of the target exception behavior from the initial teaching text; the preference identification module analyzes the second behavior data to obtain a target behavior preference; and the data adjustment module adjusts the associated teaching text according to the target abnormal behavior and the target behavior preference to obtain a target teaching text.

Description

Technical Field

[0001] The present invention relates to the field of smart education technology, and in particular to a smart education system. Background Art

[0002] In the education process, teachers usually prepare detailed lesson plans before teaching to guide and regulate teaching activities. The formulation of lesson plans is an important part of teaching preparation, which helps to ensure the clarity of teaching objectives, the systematicness of content, and the rationality of teaching methods. However, the existing lesson plan design may be too rigid, and it is difficult to adjust teaching strategies in time according to students' performance and feedback, making it difficult to effectively respond to various changes in the classroom. In other words, many lesson plans do not fully take into account students' individual differences, such as learning ability, interests, cognitive level, and cultural background. This one-size-fits-all teaching design is difficult to meet the needs of different students, resulting in some students being unable to keep up with the teaching progress or losing interest in learning, which further reduces the quality of teachers' teaching. Summary of the invention

[0003] The main purpose of the embodiments of the present invention is to provide a smart education system, which aims to solve the problem in related technologies that teachers' teaching plans are difficult to meet the needs of different students, resulting in some students being unable to keep up with the teaching progress, thereby reducing the teacher's teaching quality.

[0004] In a first aspect, an embodiment of the present invention provides a smart education system, including:

[0005] A data collection module, used to obtain first behavior data corresponding to the target students under the initial teaching text and obtain second behavior data corresponding to the target students watching the teaching video;

[0006] An abnormality analysis module, configured to perform abnormal behavior analysis on the first behavior data to obtain a first abnormal behavior corresponding to the target student and a first abnormal position corresponding to the first abnormal behavior;

[0007] A data acquisition module, configured to acquire adjacent behavior data corresponding to the target student from the first behavior data according to the first abnormal position;

[0008] A behavior prediction module, configured to perform behavior reconstruction based on the adjacent behavior data to obtain a target predicted behavior corresponding to the target student under the first abnormal behavior;

[0009] an abnormal screening module, used for screening out a second abnormal behavior corresponding to the target student according to the target predicted behavior and the first abnormal behavior;

[0010] an abnormality classification module, used to perform abnormality classification according to the second abnormal behavior to obtain an abnormality type corresponding to the second abnormal behavior;

[0011] An abnormality determination module, used to determine a target abnormal behavior corresponding to the target student according to the abnormality type, and obtain an associated teaching text corresponding to the target abnormal behavior from the initial teaching text;

[0012] A preference identification module, used for performing behavior analysis on the second behavior data to obtain a target behavior preference corresponding to the target student;

[0013] A data adjustment module is used to adjust the associated teaching text according to the target abnormal behavior and the target behavior preference to obtain the target teaching text corresponding to the target student.

[0014] The embodiment of the present invention provides a smart education system, which includes: a data acquisition module, which is used to obtain the first behavior data corresponding to the target student under the initial teaching text and obtain the second behavior data corresponding to the target student watching the teaching video, and then by collecting the behavior data of the target student in the initial teaching text and watching the teaching video, the student's learning habits and interaction mode can be fully understood, and then support for the subsequent provision of personalized teaching texts. An abnormal analysis module, which is used to perform abnormal behavior analysis on the first behavior data to obtain the first abnormal behavior corresponding to the target student and the first abnormal position corresponding to the first abnormal behavior, and then by analyzing the first abnormal behavior in the first behavior data, the target student's possible learning disabilities or comprehension difficulties can be identified, and support for the subsequent adjustment of the teaching text. The data acquisition module is used to obtain the adjacent behavior data corresponding to the target student from the first behavior data according to the first abnormal position; the behavior prediction module is used to reconstruct the behavior according to the adjacent behavior data to obtain the target predicted behavior corresponding to the target student under the first abnormal behavior; the abnormal screening module is used to screen out the second abnormal behavior corresponding to the target student according to the target predicted behavior and the first abnormal behavior; the abnormal classification module is used to classify the abnormality according to the second abnormal behavior to obtain the abnormal type corresponding to the second abnormal behavior; the abnormal determination module is used to determine the target abnormal behavior corresponding to the target student according to the abnormal type, and obtain the associated teaching text corresponding to the target abnormal behavior from the initial teaching text; the preference identification module is used to perform behavioral analysis on the second behavior data to obtain the target behavior preference corresponding to the target student, and then timely discover the learning style preference of the target student to provide direction for the subsequent adjustment of the teaching text; the data adjustment module is used to adjust the associated teaching text according to the target abnormal behavior and the target behavior preference to obtain the target teaching text corresponding to the target student. Therefore, by accurately identifying the target abnormal behavior and the target behavior preference, the pertinence of the teaching text can be significantly improved, thereby effectively supporting the subsequent learning of the target students and helping them to master knowledge more efficiently. This can not only improve the teaching quality of teachers, but also solve the problem that teachers' lesson plans in related technologies are difficult to meet the needs of different students, resulting in some students being unable to keep up with the teaching progress, thereby reducing the teaching quality of teachers. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 A schematic diagram of the module structure of a smart education system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0019] It should be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0020] The embodiment of the present invention provides a smart education system. The smart education system can be applied to a terminal device, which can be an electronic device such as a tablet computer, a laptop computer, a desktop computer, a personal digital assistant, and a wearable device. The terminal device can be a server or a server cluster.

[0021] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0022] Please refer to Figure 1 , Figure 1A schematic diagram of the module structure of a smart education system provided in an embodiment of the present invention. A smart education system 200 provided in an embodiment of the present application includes a data acquisition module 201, an abnormality analysis module 202, a data acquisition module 203, a behavior prediction module 204, an abnormality screening module 205, an abnormality classification module 206, an abnormality determination module 207, a preference identification module 208, and a data adjustment module 209, wherein the data acquisition module 201 is used to obtain the first behavior data corresponding to the target student under the initial teaching text and obtain the second behavior data corresponding to the target student watching the teaching video; the abnormality analysis module 202 is used to perform abnormal behavior analysis on the first behavior data to obtain the first abnormal behavior corresponding to the target student and the first abnormal position corresponding to the first abnormal behavior; the data acquisition module 203 is used to obtain the adjacent behavior data corresponding to the target student from the first behavior data according to the first abnormal position; the behavior prediction module 204 is used to obtain the first abnormal behavior data corresponding to the target student according to the first abnormal position; the behavior prediction module 204 is used to obtain the first abnormal behavior data corresponding to the target student according to the first abnormal position; the behavior prediction module 204 is used to obtain the first abnormal behavior data corresponding to the target student according to the first abnormal position; the behavior prediction module 204 is used to obtain the first abnormal behavior data corresponding to the target student according to the first abnormal position; the behavior prediction module 204 is used to obtain the first abnormal behavior data corresponding to the target student according to the first abnormal position; the behavior prediction module 204 is used to obtain the second ... Behavior reconstruction is performed based on the adjacent behavior data to obtain the target predicted behavior corresponding to the target student under the first abnormal behavior; an abnormal screening module 205 is used to screen out the second abnormal behavior corresponding to the target student based on the target predicted behavior and the first abnormal behavior; an abnormal classification module 206 is used to perform abnormal classification based on the second abnormal behavior to obtain the abnormal type corresponding to the second abnormal behavior; an abnormal determination module 207 is used to determine the target abnormal behavior corresponding to the target student based on the abnormal type, and obtain the associated teaching text corresponding to the target abnormal behavior from the initial teaching text; a preference identification module 208 is used to perform behavior analysis on the second behavior data to obtain the target behavior preference corresponding to the target student; a data adjustment module 209 is used to adjust the associated teaching text based on the target abnormal behavior and the target behavior preference to obtain the target teaching text corresponding to the target student.

[0023] Exemplarily, the data acquisition module 201 is used to obtain the first behavior data corresponding to the target students under the initial teaching text and to obtain the second behavior data corresponding to the target students watching the teaching video, wherein the initial teaching text can be a lesson plan prepared by the teacher according to the textbook content, or it can be a handout summarized by the teacher according to the knowledge points that need to be explained.

[0024] For example, the first behavior data of the target students when they follow the teacher to learn the initial teaching text is tracked through video sensors. The first behavior data is the video data corresponding to the target students when they listen to the teacher explaining the initial teaching text. In addition, the functions of the video platform can also be used to track the second behavior data of the target students watching other teaching videos, such as viewing type, viewing time, number of playbacks, pause points, viewing speed, etc. In addition, throughout the process, it is important to ensure the privacy and security of the data, comply with relevant data protection regulations, and obtain the consent of the target students to collect and use their first behavior data and second behavior data.

[0025] Exemplarily, the abnormal analysis module 202 determines the normal behavior posture based on historical experience, and then uses statistical methods or machine learning algorithms to identify behaviors in the first behavior data that are significantly different from the normal behavior posture benchmark. These first abnormal behaviors may include lying on the table, bowing the head for a long time, etc. Then, the corresponding frame of the first abnormal behavior in the first behavior data is obtained, thereby obtaining the time information corresponding to the corresponding frame, and then determining the time information as the first abnormal position corresponding to the first abnormal behavior.

[0026] Exemplarily, the data acquisition module 203 intercepts normal behavior data corresponding to a preset range before the first abnormal position from the first behavior data according to the first abnormal position and determines it as the adjacent behavior data corresponding to the target student under the first abnormal behavior.

[0027] Exemplarily, the behavior prediction module 204 selects a suitable behavior prediction model, common models include decision trees, random forests, support vector machines, neural networks, etc., and then uses the preprocessed historical neighboring behavior data to train the selected model. In this process, the model learns the patterns and relationships in the data, so that it can predict the target behavior based on the input neighboring behavior data, thereby obtaining the target predicted behavior corresponding to the first abnormal position of the target student under the first abnormal behavior.

[0028] Exemplarily, the abnormal screening module 205 calculates the similarity between the target predicted behavior and the first abnormal behavior to obtain a target similarity value, and then when the target similarity value is less than a preset value, the first abnormal behavior corresponding to the target similarity value is determined as the second abnormal behavior corresponding to the target student; when the target similarity value is greater than or equal to the preset value, it means that the first abnormal behavior corresponding to the target similarity value is a misjudgment, and the first abnormal behavior can be directly deleted.

[0029] Exemplarily, the anomaly classification module 206 uses machine learning algorithms such as K-nearest neighbor, decision tree, support vector machine, etc. to perform model training on historical abnormal data and known abnormal types corresponding to the historical abnormal data to obtain a target classification model, and then classifies the second abnormal behavior according to the target classification model to obtain the abnormal type corresponding to the second abnormal behavior.

[0030] Exemplarily, the abnormality determination module 207 summarizes the abnormality types corresponding to all target students to obtain the target abnormal behavior corresponding to the abnormality types corresponding to all target students when the proportion exceeds a preset value, and then obtains the second abnormal position corresponding to the target type from the first abnormal position according to the target abnormal behavior, and then obtains the content currently explained by the teacher according to the second abnormal position, and then obtains the associated teaching text corresponding to the target abnormal behavior from the initial teaching text.

[0031] Exemplarily, the preference identification module 208 defines a behavior preference index based on the second behavior data. The behavior preference index is used to help quantify and evaluate behavior preferences, such as engagement, viewing time, viewing type, etc., and then calculate the average, median, standard deviation, etc. of each behavior index to understand the distribution of the data, thereby analyzing the change trend of the second behavior data over time, discovering the behavior pattern, and then calculating the correlation between different behavior indicators, identifying the key factors affecting the behavior preference, thereby clustering the second behavior data of the students, discovering different behavior patterns and preference types, and then identifying the target behavior preference of the target students based on the analysis results.

[0032] For example, if the target students like to learn geometry-related content in mathematics, it can be found that the target behavior preferences of the target students may be content related to space and imagination.

[0033] Exemplarily, the data adjustment module 209 is used to adjust the associated teaching text according to the target abnormal behavior and the target behavior preference to obtain the target teaching text corresponding to the target student.

[0034] For example, if the target abnormal behavior is frequent distraction in class, and the target behavior preference is liking to interact and participate, then specific adjustment strategies are determined based on the target abnormal behavior and the target behavior preference. These strategies may include: increasing interactivity: increasing classroom interaction and discussion sessions; introducing fun: adding interesting cases, stories or practical activities; integrating students' interests: introducing relevant content or activities based on students' interests. For example, adding group discussions and classroom interactions to the teaching text, and adding science-related practical activities to the teaching text, such as scientific experiments and scientific competitions.

[0035] For example, according to the determined adjustment strategy, the associated teaching text is specifically adjusted to obtain the target teaching text. For example, the originally boring associated teaching text is changed to explain through experiments and practical operations to obtain the target teaching text, thereby obtaining a target teaching text that better meets the needs and preferences of students. This process helps to improve the learning interest and participation of target students, improve abnormal behaviors, and enhance learning effects.

[0036] For example, after obtaining the target teaching text, when there are students with the same situation as the target students, knowledge can be imparted according to the target teaching text, thereby improving the teacher's teaching quality and the students' learning effect of learning knowledge.

[0037] In some embodiments, the abnormal analysis module includes: a target recognition module, which is used to perform target recognition on the first behavior data to obtain target posture information corresponding to the target student in each frame; a posture clustering module, which is used to perform posture clustering on the target posture information to obtain a target clustering result corresponding to the target student; a key posture determination module, which is used to determine the key posture information corresponding to the first behavior data according to the target clustering result; an abnormal scoring module, which is used to score abnormal behaviors according to the key posture information to obtain a target abnormal score corresponding to the key posture information; an abnormal behavior determination module, which is used to determine the first abnormal behavior corresponding to the target student and the first abnormal position corresponding to the first abnormal behavior according to the target abnormal score.

[0038] Exemplarily, the abnormal analysis module includes a target recognition module, a posture clustering module, a key posture determination module, an abnormal scoring module and an abnormal behavior determination module.

[0039] Exemplarily, the target recognition module performs target recognition on the first behavior data to obtain the target posture information corresponding to the target student in each frame. For example, the first behavior data includes a video or image sequence, and each frame of the image includes an image of the target student. Then, a target detection algorithm (such as YOLO, SSD, etc.) is used to detect the target student in each frame of the image, and a posture estimation algorithm (such as OpenPose, HRNet, etc.) is used to extract the joint position information of the target student to generate the target posture information of each frame.

[0040] Exemplarily, the posture clustering module performs posture clustering on the target posture information to obtain the target clustering results corresponding to the target students. For example, features such as joint positions, angles, distances, etc. are extracted from the target posture information of each frame, and then the extracted features are clustered using a clustering algorithm (such as K-means, DBSCAN, etc.), similar postures are classified into one category, and thus the target clustering results of the target students are generated, each cluster representing a posture type.

[0041] Exemplarily, the key posture determination module determines the key posture information corresponding to the first behavior data based on the target clustering results, and then analyzes the target clustering results to identify cluster clusters with higher occurrence frequencies or significant characteristics, thereby selecting clusters with higher occurrence frequencies or significant characteristics as key posture information.

[0042] Exemplarily, the abnormal scoring module scores abnormal behavior based on the key posture information to obtain the target abnormal score corresponding to the key posture information. For example, the abnormal behavior is defined as a posture duration that is too long or an abnormal joint angle, and then a scoring standard is formulated, such as 1 point for a joint angle exceeding a certain threshold, and 2 points for a posture duration exceeding a certain threshold, so that the key posture information is scored according to the scoring standard to generate an abnormal score.

[0043] Exemplarily, the abnormal behavior determination module determines the first abnormal behavior corresponding to the target student and the first abnormal position corresponding to the first abnormal behavior according to the target abnormal score. For example, a threshold value of abnormal behavior is set, and postures exceeding the threshold value are considered abnormal behaviors. Thus, the target abnormal score is compared with the threshold value to determine which postures are the first abnormal behavior. Then, the corresponding frame of the first abnormal behavior in the first behavior data is obtained, thereby obtaining the time information corresponding to the corresponding frame, and then determining the time information as the first abnormal position corresponding to the first abnormal behavior.

[0044] In some embodiments, the posture clustering module includes: a joint recognition module, which uses the joint recognition layer of the feature recognition model to identify the target joints in the target posture information to obtain the joint positions corresponding to the target joints of the target student in the target posture information; an information calculation module, which uses the information calculation layer of the feature recognition model to calculate the relative position information corresponding to the target joints in the target posture information according to the joint positions; an emotion recognition module, which uses the emotion recognition layer of the feature recognition model to determine the target emotion information corresponding to the target student according to the relative position information and the corresponding facial information in the target posture information; an information clustering module, which uses the information clustering layer of the feature recognition model to perform posture clustering according to the relative position information and the target emotion information to obtain the target clustering result corresponding to the target student.

[0045] Exemplarily, the posture clustering module includes a joint recognition module, an information calculation module, an emotion recognition module and an information clustering module.

[0046] Exemplarily, the joint recognition module uses the joint recognition layer of the feature recognition model to identify the target joints in the target posture information, and obtains the joint positions corresponding to the target joints of the target student in the target posture information. For example, the joint recognition layer is OpenPose, which can identify the key joints of the target such as shoulders, elbows, knees, etc., and then extract the position information of the target joints from the output of the joint recognition layer to generate the coordinate data of the target joints in the target posture information, that is, obtain the joint positions corresponding to the target joints.

[0047] Exemplarily, the information calculation module uses the information calculation layer of the feature recognition model to calculate the relative position information corresponding to the target joint in the target posture information according to the joint position, for example, to calculate the relative position between the target joints according to the joint position information. For example, the position of the elbow joint relative to the shoulder joint, the position of the knee joint relative to the hip joint, etc. are calculated, so as to obtain the relative position information corresponding to the target joint.

[0048] Exemplarily, the emotion recognition module uses the emotion recognition layer of the feature recognition model to determine the target emotion information corresponding to the target student based on the relative position information and the corresponding facial information in the target posture information. For example, the facial information of the target student, such as facial expression, eye state, etc., is extracted from the target posture information, and then the emotion recognition layer (such as an emotion recognition model based on deep learning) is used to analyze the facial information and relative position information to identify the target emotion information of the target student, such as happiness, sadness, anger, etc.

[0049] Exemplarily, the information clustering module uses the information clustering layer of the feature recognition model to cluster postures according to the relative position information and the target emotion information to obtain the target clustering results corresponding to the target students. The relative position information and the target emotion information are combined into a feature vector as the input of clustering, and then the feature vector is clustered using a clustering algorithm (such as K-means, DBSCAN, etc.), similar information is classified into one category, and then the target clustering results of the target students are generated, each clustering cluster represents a posture type and its corresponding emotional state.

[0050] In some embodiments, the information clustering module includes: a center determination module, which is used to use the information clustering layer to determine the initial clustering center from the target posture information according to the relative position information and the target emotion information, and obtain the initial position information and initial emotion information corresponding to the initial clustering center; a distance calculation module, which is used to use the information clustering layer to calculate the distance information between the target posture information and the initial clustering center according to the relative position information, the target emotion information, the initial position information and the initial emotion information; an initial clustering module, which is used to cluster the target posture information according to the distance information to obtain an initial clustering result; a fluctuation calculation module, which is used to perform data distribution fluctuation calculation on each first cluster in the initial clustering result to obtain the target fluctuation value corresponding to the first cluster; a cluster adjustment module, which is used to adjust the initial clustering result according to the target fluctuation value to obtain the target clustering result corresponding to the target student; wherein the distance information is obtained according to the following formula:

[0051] ;

[0052] in, Represents the distance information between the i-th target posture information and the j-th initial cluster center; represents weight information; n represents the number corresponding to the relative position information; Represents the weight information corresponding to the k-th relative position information; The kth relative position information representing the i-th target posture information; Represents the kth initial position information of the jth initial cluster center; Indicates the target emotion information corresponding to the i-th target posture information; Represents the initial emotion information corresponding to the j-th initial cluster center.

[0053] Exemplarily, the information clustering module includes a center determination module, a distance calculation module, an initial clustering module, a fluctuation calculation module and a clustering adjustment module.

[0054] Exemplarily, the center determination module uses the information clustering layer to extract relative position information and target emotion information from the target posture information as input features for clustering. Using the initial center selection method (such as random selection or selection based on some heuristics) of the clustering algorithm (such as the K-means algorithm), several initial cluster centers are selected from the target posture information, and then the relative position information and target emotion information of each initial cluster center are extracted to generate initial position information and initial emotion information.

[0055] Exemplarily, the distance calculation module calculates the distance information between each sample in the target posture information and the initial cluster center according to the following formula:

[0056] ;

[0057] in, Represents the distance information between the i-th target posture information and the j-th initial cluster center; represents weight information; n represents the number corresponding to the relative position information; Indicates the weight information corresponding to the kth relative position information; Represents the kth relative position information of the i-th target posture information; Represents the kth initial position information of the jth initial cluster center; Indicates the target emotion information corresponding to the i-th target posture information; Represents the initial emotion information corresponding to the jth initial cluster center.

[0058] Exemplarily, the formula allows the relative position information and the target emotion information to be simultaneously included in the distance calculation. By fusing features of multiple dimensions (such as the relative position of the body and the emotional state), the similarity between the target posture information and the initial cluster center can be more comprehensively measured. This fusion of multi-dimensional features can improve the accuracy and robustness of the clustering results, because single-dimensional features (such as only considering body position) may not fully reflect the complex relationship between posture and emotion. The formula can comprehensively consider the feature differences in multiple dimensions. Traditional distance measurement methods (such as Euclidean distance) may only focus on position information, while ignoring emotional information or other important features. Through the weighted summation method in the formula, the distance information can more accurately reflect the difference between the target posture information and the cluster center.

[0059] Exemplarily, the initial clustering module assigns each target posture information sample to the cluster where the nearest initial clustering center is located according to the distance information, and then sorts the assigned samples by cluster to generate an initial clustering result.

[0060] Exemplarily, the fluctuation calculation module performs data distribution fluctuation calculation on each first cluster in the initial clustering result to obtain a target fluctuation value corresponding to the first cluster. For example, the target fluctuation value is used to measure the distribution of data in each cluster. The fluctuation value can be based on statistics such as the variance and standard deviation of the samples in the cluster, and then the fluctuation calculation is performed on the samples in each first cluster to generate a target fluctuation value corresponding to each cluster.

[0061] Exemplarily, the cluster adjustment module determines whether the distribution of clusters is reasonable based on the target fluctuation value. If the target fluctuation value of a cluster is too large, it means that the data distribution within the cluster is uneven and may need to be adjusted. For clusters with too large target fluctuation values, a split operation (dividing the cluster into multiple sub-clusters) or a merge operation (merging the cluster with other clusters) can be adopted to optimize the clustering results, and then after adjustment, the target clustering results corresponding to the target students are generated.

[0062] Exemplarily, an information clustering module is implemented, including center determination, distance calculation, initial clustering, fluctuation calculation and clustering adjustment. This process helps to cluster the target posture information and adjust the clustering results according to the data distribution fluctuation, and finally obtain the target clustering results corresponding to the target students. This process can effectively identify the common posture types of target students and their corresponding emotional states, provide teachers with targeted intervention suggestions, and improve students' classroom performance and learning effects.

[0063] In some embodiments, the behavior prediction module includes: using the reconstruction behavior layer of the behavior prediction model to use an autoencoder combined with the neighboring behavior data to perform behavior reconstruction to obtain a first predicted behavior; using the prediction behavior layer of the behavior prediction model to use an encoder and a decoder to combine the neighboring behavior data to perform behavior prediction to obtain a second predicted behavior; using the behavior fusion layer of the behavior prediction model to perform behavior fusion based on the first predicted behavior and the second predicted behavior to obtain the target predicted behavior corresponding to the target student under the first abnormal behavior.

[0064] Exemplarily, the reconstruction behavior layer uses an autoencoder combined with neighboring behavior data to reconstruct the behavior and obtain the first predicted behavior. For example, collect neighboring behavior data of the target student, which includes but is not limited to posture information, action sequence, etc., and then build an autoencoder model. The autoencoder is usually composed of an encoder and a decoder. The encoder compresses the input data into a low-dimensional representation, and the decoder reconstructs the low-dimensional representation into the original data, and then pre-processes the neighboring behavior data to make it suitable for the input format of the autoencoder, such as normalization, padding, etc. Thus, the neighboring behavior data is reconstructed using the autoencoder, that is, the neighboring behavior data is compressed into a low-dimensional representation by the encoder, and then the low-dimensional representation is reconstructed into the first predicted behavior by the decoder.

[0065] Exemplarily, the prediction behavior layer of the behavior prediction model uses an encoder and a decoder to combine neighboring behavior data for behavior prediction, and constructs an encoder and a decoder model. The encoder compresses the input data into a low-dimensional representation, and the decoder converts the low-dimensional representation into the future predicted behavior. The neighboring behavior data is preprocessed to make it suitable for the input format of the encoder and the decoder. The encoder is used to encode the neighboring behavior data to generate a low-dimensional representation, and then the decoder is used to convert the low-dimensional representation into the second predicted behavior in the future.

[0066] Exemplarily, the behavior fusion layer of the behavior prediction model is used to perform behavior fusion according to the first predicted behavior and the second predicted behavior to obtain the target predicted behavior corresponding to the target student under the first abnormal behavior. For example, if weighted average is used, it is necessary to determine the weights of the first predicted behavior and the second predicted behavior, and then pre-process the first predicted behavior and the second predicted behavior to ensure that they have the same dimension and format, and then fuse the first predicted behavior and the second predicted behavior according to the fusion rule to generate the target predicted behavior corresponding to the target student under the first abnormal behavior.

[0067] For example, weighted average is selected as the behavior fusion method. The weights of the first predicted behavior and the second predicted behavior are determined. For example, the weight of the first predicted behavior is 0.6, and the weight of the second predicted behavior is 0.4. The first predicted behavior and the second predicted behavior are then normalized to ensure that they have the same dimension and format. The first predicted behavior and the second predicted behavior are then fused using the weighted average method to generate the target predicted behavior corresponding to the target student under the first abnormal behavior.

[0068] In some embodiments, the abnormal screening module includes: a similarity calculation module, used to calculate the similarity between the target predicted behavior and the first abnormal behavior; and a screening processing module, used to screen out the second abnormal behavior corresponding to the target student from the first abnormal behavior according to the similarity.

[0069] Exemplarily, the similarity calculation module represents the target predicted behavior and the first abnormal behavior as comparable feature vectors. These features may include key point positions of postures, sequences of actions, emotional states, etc., and then calculates the similarity between the target predicted behavior and each first abnormal behavior using similarity measurement methods such as cosine similarity, Euclidean distance, correlation coefficient, etc.

[0070] Exemplarily, the screening processing module sets a similarity threshold to determine which first abnormal behaviors are sufficiently similar to the target predicted behaviors, and then screens out the first abnormal behaviors whose similarity is lower than the threshold according to the set threshold, and then uses the screened behaviors as the second abnormal behaviors of the target students.

[0071] In some embodiments, the abnormal classification module includes: a first feature extraction module, which is used to perform deflection feature extraction on the second abnormal behavior to obtain the first behavior feature corresponding to the second abnormal behavior; a second feature extraction module, which is used to perform gradient feature extraction on the second behavior feature using adjacent information to obtain the second behavior feature corresponding to the second abnormal behavior under instantaneous behavior; a third feature extraction module, which is used to perform emotion recognition on the second abnormal behavior to obtain the third behavior feature corresponding to the second abnormal behavior; a feature fusion module, which is used to fuse the first behavior feature, the second behavior feature and the third behavior feature to obtain the target behavior feature corresponding to the second abnormal behavior; a correlation calculation module, which is used to determine a preset type and a preset feature variable corresponding to the preset type, and determine the correlation value between the second abnormal behavior and the preset type according to the target behavior feature and the preset feature variable; a type determination module, which is used to determine the abnormal type corresponding to the second abnormal behavior from the preset type according to the correlation value; wherein the correlation value is obtained according to the following formula:

[0072]

[0073] in, represents the correlation value between the i-th second abnormal behavior and the j-th preset type; represents the preset characteristic variable corresponding to the j-th preset type, represents the target behavior feature corresponding to the i-th second abnormal behavior; represents the balance parameter; r represents the number of corresponding features in the target behavior feature; err represents the error parameter; represents the similarity between the preset feature variable corresponding to the j-th preset type and the target behavior feature corresponding to the i-th second abnormal behavior; represents the adjustment parameter; Represents a preset feature vector corresponding to the wth adjacent preset type corresponding to the jth preset type.

[0074] Exemplarily, the first feature extraction module extracts the deflection feature of the second abnormal behavior, obtains the first behavior feature corresponding to the second abnormal behavior, and analyzes the deflection in the second abnormal behavior (such as a sudden change in body posture, abnormal deviation of movement, etc.). The deflection feature can be captured by the relative change of posture data. The extracted deflection feature is represented as a feature vector, such as a deflection angle, a deflection amplitude, etc. For example, the deflection angle of the body posture in the second abnormal behavior (such as a sudden raising of the arm or an abnormal bending of the leg) is extracted, and then the deflection angle and amplitude are recorded to form the first behavior feature.

[0075] Exemplarily, the second feature extraction module uses adjacent information to perform gradient feature extraction on the second behavior feature to obtain the second behavior feature corresponding to the instantaneous behavior of the second abnormal behavior. The front and rear frame information of the second abnormal behavior is analyzed to capture the change trend and instantaneous gradient features of its behavior, and then the gradient features of the second abnormal behavior are calculated, such as the rate and direction of the behavior change. For example, the front and rear frames of the second abnormal behavior (such as frames with a time interval of 0.5 seconds) are analyzed, and then the gradient features of the behavior change, such as the speed and direction of the action change, are calculated to form the second behavior feature.

[0076] Exemplarily, the third feature extraction module performs emotion recognition on the second abnormal behavior to obtain a third behavior feature corresponding to the second abnormal behavior. For example, an emotion recognition model or rule (such as facial expression, voice tone, behavior pattern, etc.) is used to identify the emotional state in the second abnormal behavior. The recognized emotional state is represented as a feature vector, such as an emotion category (anxiety, anger, etc.) and an emotion intensity, so that the emotional state (such as anxiety) in the second abnormal behavior is identified through facial expressions and behavior patterns, and the emotion category and intensity are recorded to form a third behavior feature.

[0077] Exemplarily, the feature fusion module combines the first behavior feature, the second behavior feature and the third behavior feature through a feature fusion method, such as weighted averaging, concatenation, neural network fusion, etc., to obtain a target behavior feature corresponding to the second abnormal behavior.

[0078] Exemplarily, the correlation calculation module determines the preset type and the preset characteristic variable corresponding to the preset type, and determines the correlation value between the second abnormal behavior and the preset type according to the target behavior feature and the preset characteristic variable. A corresponding characteristic variable is defined for each preset abnormal type (such as anxiety, happiness, etc.), and then the correlation value between the target behavior feature and each preset type characteristic variable is calculated according to the following formula:

[0079]

[0080] in, represents the correlation value between the i-th second abnormal behavior and the j-th preset type; represents the preset feature variable corresponding to the j-th preset type, represents the target behavior feature corresponding to the i-th second abnormal behavior; represents the balance parameter; r represents the number of corresponding features in the target behavior feature; err represents the error parameter; Represents the similarity between the preset feature variable corresponding to the j-th preset type and the target behavior feature corresponding to the i-th second abnormal behavior; represents the adjustment parameter; Represents the preset feature vector corresponding to the wth adjacent preset type corresponding to the jth preset type.

[0081] Exemplarily, the balance parameter helps to balance the importance of different features and ensure that no feature overly dominates the calculation of the correlation value. The error parameter allows the model to tolerate imperfections in feature matching to a certain extent, making the correlation calculation more robust, and then by quantifying the similarity between the preset feature variables and the target behavior features, the degree of match between the second abnormal behavior and the preset type can be objectively measured. The adjacent preset types take into account the influence of adjacent preset types, which helps to more accurately determine the type of abnormal behavior when there are fuzzy boundaries between types. Therefore, through the above-mentioned comprehensive calculation, the second abnormal behavior can be more accurately classified into the most appropriate preset type, thereby improving the accuracy of abnormal behavior identification, and then the relationship between the second abnormal behavior and the preset type can be comprehensively and accurately evaluated, thereby improving the overall performance and reliability of abnormal behavior classification.

[0082] Exemplarily, the type determination module determines the abnormal type corresponding to the second abnormal behavior from the preset types according to the correlation value. A correlation threshold is set for each preset type, and the type with a correlation value exceeding the threshold can be considered as a candidate type, and then the preset type with the highest correlation value is selected as the abnormal type of the second abnormal behavior.

[0083] In some embodiments, the preference identification module includes: a first analysis module, which is used to perform viewing behavior analysis on the second behavior data to obtain a first viewing behavior sequence and a second viewing behavior sequence corresponding to the target student, wherein the first viewing behavior sequence is used to characterize the sequence corresponding to the target student watching the video completely, and the second viewing behavior sequence is used to characterize the sequence corresponding to the target student not watching the video completely; a second analysis module, which is used to perform viewing result analysis on the second behavior data to obtain a first viewing score sequence and a second viewing score sequence corresponding to the target student, wherein the first viewing score sequence is used to characterize the sequence corresponding to the target student liking to watch the video, and the second viewing score sequence is used to characterize the sequence corresponding to the target student not liking to watch the video; a first embedding processing module, which is used to perform feature characterization on the first viewing behavior sequence based on the first embedding layer of the interest recognition model to obtain a first feature vector and to perform feature characterization on the first viewing score sequence to obtain a second feature vector, wherein the interest recognition model is obtained by taking the first historical behavior sequence, the second historical behavior sequence, the first historical score sequence, and the second historical score sequence corresponding to the history student as model inputs, and taking the historical behavior preferences corresponding to the history student under the first historical behavior sequence, the second historical behavior sequence, the first historical score sequence, and the second historical score sequence as model outputs for model training; wherein the first The historical behavior sequence is used to characterize the sequence corresponding to the history student watching the video completely, the second historical behavior sequence is used to characterize the sequence corresponding to the history student not watching the video completely, the first historical score sequence is used to characterize the sequence corresponding to the history student liking to watch the video, and the second historical score sequence is used to characterize the sequence corresponding to the history student not liking to watch the video; the second embedding processing module is used to characterize the second viewing behavior sequence based on the second embedding layer of the interest recognition model to obtain a third feature vector and to characterize the second viewing score sequence to obtain a fourth feature vector; the first denoising module is used to denoise the first feature vector based on the first interest denoising layer of the interest recognition model The first denoised vector is obtained by denoising the second feature vector and the second denoised vector is obtained by denoising the second feature vector; the second denoising module is used to denoise the third feature vector based on the second interest denoising layer of the interest recognition model to obtain a third denoised vector and to denoise the fourth feature vector to obtain a fourth denoised vector; the vector fusion module is used to perform feature fusion on the first denoised vector, the second denoised vector, the third denoised vector and the fourth denoised vector based on the multi-head attention layer of the interest recognition model to obtain a target fusion vector; the preference classification module is used to perform multi-target preference classification according to the target fusion vector based on the preference classification layer of the interest recognition model to obtain the target behavior preference corresponding to the target student.

[0084] Exemplarily, before using the interest recognition model to analyze the behavior preferences of the target students, it is first necessary to construct a comprehensive training data set. Specifically, it is necessary to collect the first historical behavior sequence and the second historical behavior sequence of multiple history students, wherein the first historical behavior sequence records the behavior data of the history students who have watched the video in full, and the second historical behavior sequence records the behavior data of the history students who have not watched the video in full. By interactively analyzing the first historical behavior sequence and the second historical behavior sequence, such as likes, comments, and shares, the videos that the history students like are identified to generate the first historical score sequence and the videos that the history students do not like are identified to generate the second historical score sequence. Next, by manual annotation, the first historical behavior sequence, the second historical behavior sequence, the first historical score sequence, and the second historical score sequence are combined to perform preference annotation on the behavior preferences of the history students so as to obtain the historical behavior preferences corresponding to the historical users. The historical behavior preferences are used to characterize the behavior preference types corresponding to the history students under the first historical behavior sequence, the second historical behavior sequence, the first historical score sequence, and the second historical score sequence.

[0085] Exemplarily, the first historical behavior sequence, the second historical behavior sequence, the first historical score sequence, and the second historical score sequence corresponding to the historical user are used as the input of the interest recognition model, and the historical behavior preferences corresponding to the historical user are used as the output of the interest recognition model. Multiple rounds of model training are performed with cross entropy as the loss function, and finally an interest recognition model that can accurately identify students' behavioral preferences is obtained.

[0086] Exemplarily, the first analysis module processes the second behavior data (i.e., the viewing behavior data of the target students) to obtain videos that the target students have watched completely, which are recorded as the first viewing behavior sequence, and obtains videos that the target students have not watched completely, which are recorded as the second viewing behavior sequence.

[0087] Exemplarily, the second analysis module performs interactive analysis on the second behavior data to analyze the videos that the target students like based on the viewing results (such as likes, comments, shares, etc.) to form a first viewing score sequence and the videos that the target students do not like to form a second viewing score sequence.

[0088] Exemplarily, the interest recognition model includes a first embedding layer, a second embedding layer, a first interest denoising layer, a second interest denoising layer, a multi-head attention layer, and a preference classification layer. The first embedding processing module uses the first embedding layer of the interest recognition model to extract features from the first viewing behavior sequence to obtain a first feature vector, and uses the same embedding layer to extract features from the first viewing score sequence to obtain a second feature vector.

[0089] Exemplarily, the second embedding processing module uses the second embedding layer of the interest recognition model to perform feature extraction on the second viewing behavior sequence to obtain a third feature vector, and uses the same embedding layer to perform feature extraction on the second viewing score sequence to obtain a fourth feature vector.

[0090] Exemplarily, the first denoising module uses the first interest denoising layer of the interest recognition model to denoise the first feature vector to obtain a first denoised vector, and performs the same denoising process on the second feature vector to obtain a second denoised vector.

[0091] Exemplarily, the second denoising module uses the second interest denoising layer of the interest recognition model to denoise the third feature vector to obtain a third denoised vector. The fourth feature vector is denoised in the same manner to obtain a fourth denoised vector.

[0092] Exemplarily, the first denoised vector, the second denoised vector, the third denoised vector and the fourth denoised vector of the multi-head attention layer of the interest recognition model are used for feature fusion to obtain a target fusion vector. Then, the target fusion vector is classified using the preference classification layer of the interest recognition model, for example, the target behavior preference of the target student is determined by a classification algorithm (such as softmax, etc.).

[0093] In some embodiments, the data adjustment module includes: a preference scoring module, which is used to summarize all the target behavior preferences to obtain an overall behavior preference, and obtain a target preference score corresponding to the target behavior preference according to the target proportion of the target behavior preference in the overall behavior preference; a preference sorting module, which is used to sort the target behavior preference according to the target preference score to obtain a target preference sorting corresponding to the target student; a rejection determination module, which is used to determine the target rejection degree of the target student to the associated teaching text according to the target abnormal behavior; and a text generation module, which is used to obtain the target teaching text corresponding to the target student by using a text generation model in combination with the associated teaching text according to the target preference sorting and the target rejection degree.

[0094] Exemplarily, the data adjustment module includes a preference scoring module, a preference sorting module, an exclusion determination module and a text generation module.

[0095] Exemplarily, the preference scoring module summarizes all target behavior preferences to obtain an overall behavior preference, and obtains a target preference score according to the target proportion of the target behavior preference in the overall behavior preference.

[0096] Exemplarily, the preference sorting module sorts the target behavior preferences according to the target preference scores to obtain the target preference sorting corresponding to the target students. For example, the target behavior preferences are sorted from high to low according to the target preference scores to obtain the target preference sorting corresponding to the target students.

[0097] Exemplarily, the rejection determination module determines the target rejection degree of the target student to the associated teaching text according to the target abnormal behavior. For example, by analyzing the target abnormal behavior, its relevance to the associated teaching text is determined. For example, if the target abnormal behavior is "inattention", it may indicate that the student is not interested in or feels repulsive to the current teaching content. Thus, according to the severity of the target abnormal behavior, the rejection degree of the target student to the associated teaching text is determined. For example, the rejection degree can be divided into three levels: low, medium, and high.

[0098] Exemplarily, the text generation module uses the text generation model in combination with the associated teaching text according to the target preference ranking and the target rejection degree to obtain the target teaching text corresponding to the target student. For example, according to the target preference ranking, priority is given to the content that students prefer. Considering the target rejection degree, the content and presentation of the teaching text are adjusted to reduce the students' sense of rejection. For example, if the student does not like the text and has a high rejection degree, the complexity of the text content can be reduced, and simpler language and more diagrams can be used. Thus, the integrated preference ranking and rejection degree information are used to generate teaching content that meets the student's preferences by using the text generation model and combining the associated teaching text. For example, a natural language generation (NLG) model can be used to generate a more personalized and more acceptable teaching text according to the student's preferences and rejection degree, so as to obtain the target teaching text corresponding to the target student.

[0099] In some embodiments, the text generation module includes: a knowledge identification module, which is used to identify knowledge points of the associated teaching text to obtain target knowledge points corresponding to the associated teaching text; a path identification module, which is used to determine the wrong explanation path corresponding to the target knowledge point according to the target exclusion degree, and obtain the first correct explanation path corresponding to the target knowledge point according to the target preference sorting; a path trimming module, which is used to trim the wrong explanation path involved in the first correct explanation path to obtain the second correct explanation path corresponding to the target knowledge point; a knowledge generation module, which is used to use the text generation model to obtain the correct teaching text corresponding to the associated teaching text according to the second correct explanation path and the target knowledge point; a text fusion module, which is used to fuse the initial teaching text and the correct teaching text to obtain the target teaching text corresponding to the target student.

[0100] Exemplarily, the text generation module includes a knowledge identification module, a path identification module, a path repair module, a knowledge generation module and a text fusion module.

[0101] Exemplarily, the knowledge recognition module uses natural language processing technology (such as named entity recognition, keyword extraction, etc.) to extract key knowledge points from the associated teaching text. For example, if the associated teaching text is about "Newton's Second Law", the extracted key knowledge points may include "Newton's Second Law", "force", "mass", "acceleration", etc., thereby obtaining the target knowledge points corresponding to the associated teaching text.

[0102] Exemplarily, the path identification module determines which explanation paths may cause rejection or misunderstanding of the target students according to the degree of target rejection. For example, if the target rejection degree is high, it may be necessary to avoid complex mathematical derivations or formalized language. Then, according to the target preference sorting, the explanation path that best meets the student's preference is selected. For example, if the student prefers simple and intuitive explanations, the explanation path containing more diagrams and examples is selected. Thus, the wrong explanation path and the first correct explanation path corresponding to the target knowledge point are obtained.

[0103] Exemplarily, the path trimming module checks whether the first correct explanation path contains any wrong explanation paths. If so, it is trimmed to ensure that the explanation path meets the student's preferences and rejection level. For example, if the first correct explanation path contains complex mathematical derivations, it can be replaced with simpler explanations or diagrams to obtain a second correct explanation path corresponding to the target knowledge point.

[0104] Exemplarily, the knowledge generation module generates correct teaching text according to the second correct explanation path and target knowledge points using a text generation model (such as a deep learning model, a natural language generation model, etc.). For example, an NLG model can be used to generate text containing diagrams, examples, and simple explanations to make it easier to understand.

[0105] Exemplarily, the text fusion module fuses the correct teaching text with the initial teaching text to ensure the coherence and integrity of the teaching content. For example, the correct explanation path can be inserted or replaced in the initial teaching text to make it more in line with the needs of students, thereby obtaining the target teaching text corresponding to the target students.

[0106] An embodiment of the present invention also provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any smart education system provided in the description of the embodiment of the present invention.

[0107] The storage medium may be an internal storage unit of the terminal device in the aforementioned embodiment, such as a hard disk or memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal device.

[0108] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware embodiment, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0109] It should be understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0110] The serial numbers of the embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A smart education system, characterized in that: The system comprises: A data collection module, used to obtain first behavior data corresponding to the target students under the initial teaching text and obtain second behavior data corresponding to the target students watching the teaching video; An abnormality analysis module, configured to perform abnormal behavior analysis on the first behavior data to obtain a first abnormal behavior corresponding to the target student and a first abnormal position corresponding to the first abnormal behavior; A data acquisition module, configured to acquire adjacent behavior data corresponding to the target student from the first behavior data according to the first abnormal position; A behavior prediction module, configured to perform behavior reconstruction based on the adjacent behavior data to obtain a target predicted behavior corresponding to the target student under the first abnormal behavior; an abnormal screening module, used for screening out a second abnormal behavior corresponding to the target student according to the target predicted behavior and the first abnormal behavior; an abnormality classification module, used to perform abnormality classification according to the second abnormal behavior to obtain an abnormality type corresponding to the second abnormal behavior; An abnormality determination module, used to determine a target abnormal behavior corresponding to the target student according to the abnormality type, and obtain an associated teaching text corresponding to the target abnormal behavior from the initial teaching text; A preference identification module, used for performing behavior analysis on the second behavior data to obtain a target behavior preference corresponding to the target student; A data adjustment module is used to adjust the associated teaching text according to the target abnormal behavior and the target behavior preference to obtain the target teaching text corresponding to the target student.

2. The system according to claim 1, characterized in that The abnormality analysis module comprises: A target recognition module, used for performing target recognition on the first behavior data to obtain target posture information corresponding to the target student in each frame; A posture clustering module, used for performing posture clustering on the target posture information to obtain a target clustering result corresponding to the target student; A key posture determination module, used to determine key posture information corresponding to the first behavior data according to the target clustering result; An abnormality scoring module, used to score abnormal behavior according to the key posture information, and obtain a target abnormality score corresponding to the key posture information; An abnormal behavior determination module is used to determine the first abnormal behavior corresponding to the target student and the first abnormal position corresponding to the first abnormal behavior according to the target abnormal score.

3. The system according to claim 2, characterized in that The posture clustering module comprises: A joint recognition module, used to recognize the target joint in the target posture information by using the joint recognition layer of the feature recognition model to obtain the joint position corresponding to the target joint of the target student in the target posture information; An information calculation module, used for calculating the relative position information corresponding to the target joint in the target posture information according to the joint position by using the information calculation layer of the feature recognition model; An emotion recognition module, used to determine the target emotion information corresponding to the target student according to the relative position information and the corresponding facial information in the target posture information using the emotion recognition layer of the feature recognition model; An information clustering module is used to utilize the information clustering layer of the feature recognition model to perform posture clustering according to the relative position information and the target emotion information to obtain a target clustering result corresponding to the target student.

4. The system according to claim 3, characterized in that The information clustering module comprises: A center determination module, used to determine an initial clustering center from the target posture information according to the relative position information and the target emotion information using the information clustering layer, and obtain initial position information and initial emotion information corresponding to the initial clustering center; A distance calculation module, used to calculate the distance information between the target posture information and the initial clustering center according to the relative position information, the target emotion information, the initial position information and the initial emotion information by using the information clustering layer; An initial clustering module, used for clustering the target posture information according to the distance information to obtain an initial clustering result; A fluctuation calculation module, used for performing data distribution fluctuation calculation on each first cluster in the initial clustering result to obtain a target fluctuation value corresponding to the first cluster; A clustering adjustment module, used for adjusting the initial clustering result according to the target fluctuation value to obtain the target clustering result corresponding to the target student; The distance information is obtained according to the following formula: ； in, Represents the distance information between the i-th target posture information and the j-th initial cluster center; represents weight information; n represents the number corresponding to the relative position information; Represents the weight information corresponding to the k-th relative position information; The kth relative position information representing the i-th target posture information; Represents the kth initial position information of the jth initial cluster center; Indicates the target emotion information corresponding to the i-th target posture information; Represents the initial emotion information corresponding to the j-th initial cluster center.

5. The system according to claim 1, characterized in that The behavior prediction module comprises: Using the reconstructed behavior layer of the behavior prediction model and the autoencoder in combination with the neighboring behavior data to perform behavior reconstruction to obtain a first predicted behavior; Using the predicted behavior layer of the behavior prediction model to collect the neighboring behavior data using an encoder and a decoder to perform behavior prediction to obtain a second predicted behavior; The behavior fusion layer of the behavior prediction model is used to perform behavior fusion according to the first predicted behavior and the second predicted behavior to obtain the target predicted behavior corresponding to the target student under the first abnormal behavior.

6. The system according to claim 1, characterized in that The abnormal screening module comprises: A similarity calculation module, used to calculate the similarity between the target predicted behavior and the first abnormal behavior; A screening processing module is used to screen out the second abnormal behavior corresponding to the target student from the first abnormal behaviors according to the similarity.

7. The system according to claim 1, characterized in that The abnormal classification module comprises: A first feature extraction module, configured to perform deflection feature extraction on the second abnormal behavior to obtain a first behavior feature corresponding to the second abnormal behavior; A second feature extraction module, configured to perform gradient feature extraction on the second behavior feature by using adjacent information to obtain a second behavior feature corresponding to the second abnormal behavior under instantaneous behavior; a third feature extraction module, configured to perform emotion recognition on the second abnormal behavior to obtain a third behavior feature corresponding to the second abnormal behavior; A feature fusion module, configured to fuse the first behavior feature, the second behavior feature, and the third behavior feature to obtain a target behavior feature corresponding to the second abnormal behavior; a correlation calculation module, used to determine a preset type and a preset characteristic variable corresponding to the preset type, and determine a correlation value between the second abnormal behavior and the preset type according to the target behavior feature and the preset characteristic variable; a type determination module, configured to determine the abnormal type corresponding to the second abnormal behavior from the preset types according to the correlation value; The correlation value is obtained according to the following formula: ； in, represents the correlation value between the i-th second abnormal behavior and the j-th preset type; represents the preset characteristic variable corresponding to the j-th preset type, represents the target behavior feature corresponding to the i-th second abnormal behavior; represents the balance parameter; r represents the number of corresponding features in the target behavior feature; err represents the error parameter; represents the similarity between the preset feature variable corresponding to the j-th preset type and the target behavior feature corresponding to the i-th second abnormal behavior; represents the adjustment parameter; Represents a preset feature vector corresponding to the wth adjacent preset type corresponding to the jth preset type.

8. The system according to claim 1, characterized in that The preference identification module comprises: A first analysis module is configured to perform viewing behavior analysis on the second behavior data to obtain a first viewing behavior sequence and a second viewing behavior sequence corresponding to the target student, wherein the first viewing behavior sequence is used to represent a sequence corresponding to when the target student has watched the video completely, and the second viewing behavior sequence is used to represent a sequence corresponding to when the target student has not watched the video completely; A second analysis module is configured to perform viewing result analysis on the second behavior data to obtain a first viewing score sequence and a second viewing score sequence corresponding to the target student, wherein the first viewing score sequence is used to represent a sequence corresponding to the target student's liking to watch videos, and the second viewing score sequence is used to represent a sequence corresponding to the target student's disliking to watch videos; A first embedding processing module is used to characterize the first viewing behavior sequence based on the first embedding layer of the interest recognition model to obtain a first feature vector and to characterize the first viewing score sequence to obtain a second feature vector, wherein the interest recognition model is obtained by taking the first historical behavior sequence, the second historical behavior sequence, the first historical score sequence, and the second historical score sequence corresponding to the history student as model inputs, and taking the historical behavior preferences of the history student under the first historical behavior sequence, the second historical behavior sequence, the first historical score sequence, and the second historical score sequence as model outputs for model training; wherein the first historical behavior sequence is used to characterize the sequence corresponding to the video that the history student has watched completely, the second historical behavior sequence is used to characterize the sequence corresponding to the video that the history student has not watched completely, the first historical score sequence is used to characterize the sequence corresponding to the video that the history student likes to watch, and the second historical score sequence is used to characterize the sequence corresponding to the video that the history student does not like to watch; A second embedding processing module, configured to perform feature characterization on the second viewing behavior sequence based on the second embedding layer of the interest recognition model to obtain a third feature vector and to perform feature characterization on the second viewing score sequence to obtain a fourth feature vector; A first denoising module, configured to perform denoising processing on the first feature vector based on the first interest denoising layer of the interest recognition model to obtain a first denoised vector and to perform denoising processing on the second feature vector to obtain a second denoised vector; A second denoising module, configured to perform denoising processing on the third feature vector based on the second interest denoising layer of the interest recognition model to obtain a third denoised vector and to perform denoising processing on the fourth feature vector to obtain a fourth denoised vector; A vector fusion module, configured to perform feature fusion on the first denoised vector, the second denoised vector, the third denoised vector and the fourth denoised vector based on the multi-head attention layer of the interest recognition model to obtain a target fused vector; A preference classification module is used to perform multi-target preference classification based on the preference classification layer of the interest recognition model according to the target fusion vector to obtain the target behavior preference corresponding to the target student.

9. The system according to claim 1, characterized in that The data adjustment module comprises: A preference scoring module, used to summarize all the target behavior preferences to obtain an overall behavior preference, and obtain a target preference score corresponding to the target behavior preference according to the target proportion of the target behavior preference in the overall behavior preference; A preference sorting module, used to sort the target behavior preferences according to the target preference scores to obtain the target preference sorting corresponding to the target students; A rejection determination module, used to determine the target rejection degree of the target student to the associated teaching text according to the target abnormal behavior; A text generation module is used to obtain the target teaching text corresponding to the target student by using a text generation model in combination with the associated teaching text according to the target preference ranking and the target exclusion degree.

10. The system according to claim 9, characterized in that The text generation module comprises: A knowledge recognition module, used for performing knowledge point recognition on the associated teaching text to obtain target knowledge points corresponding to the associated teaching text; A path identification module, used to determine an incorrect explanation path corresponding to the target knowledge point according to the target rejection degree, and obtain a first correct explanation path corresponding to the target knowledge point according to the target preference ranking; A path trimming module, used for trimming the erroneous explanation path involved in the first correct explanation path to obtain a second correct explanation path corresponding to the target knowledge point; A knowledge generation module, configured to obtain a correct teaching text corresponding to the associated teaching text according to the second correct explanation path and the target knowledge point by using the text generation model; The text fusion module is used to fuse the initial teaching text and the correct teaching text to obtain the target teaching text corresponding to the target student.

Images