Student information management system

Abstract

The invention relates to the technical field of information management, and discloses a student information management system, which comprises an information acquisition module for acquiring student behavior data through a multi-source device, the student behavior data including personal identities, learning activities, classroom behaviors and consumption tracks; the secure transmission module is used for carrying out segmented encryption on the data by adopting a dynamic encryption strategy based on a national cryptographic algorithm; the information processing module is used for decrypting the data through a trusted execution environment, extracting spatio-temporal features and cross-modal association rules by using a multi-modal fusion analysis model, and generating behavior feature vectors and a rule set; the behavior portrait module is combined with an improved K-means algorithm to dynamically generate quantitative labels including learning concentration degree, activity participation degree and consumption tendency; and the visualization module is used for outputting the portrait data to a management terminal through an encryption API interface, and realizing hierarchical access control based on the role permission matrix. The intelligent level of student behavior management can be improved, and scientific support is provided for education management decision.

Description

Technical Field

[0001] The present invention relates to the technical field of information management, and particularly relates to a student information management system. Background Art

[0002] In the field of student information management, with the construction of smart campuses, it has become a trend to conduct refined management of students' specific information. Traditional student information management systems usually cover the management of various aspects of student grades, course selection, attendance, etc., but there are still limitations in the management of a specific piece of information. For example, the management of students' behavior data (such as learning data, activity data, consumption data) has not been fully explored and utilized. In addition, existing systems still need to be further optimized in terms of data collection, analysis, and processing, especially in the real-time performance, accuracy, and intelligence level of dynamic data.

[0003] Therefore, there is an urgent need to provide a technical solution to solve the above problems. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a student information management system, with the following technical solutions: A student information management system includes: an information collection module, a secure transmission module, an information processing module, a behavior portrait module, and a visualization module; The information collection module is used to: collect students' behavior data in real time through multi-source sensors and mobile terminal devices deployed in campus scenarios, and the students' behavior data includes personal identity identifiers, learning activity data, classroom behavior data, and consumption trajectory data; The secure transmission module is used to: establish a two-way authentication channel between the information collection module and the information processing module, and perform segmented encryption on the students' behavior data using a dynamic encryption strategy based on national cryptographic algorithms to obtain encrypted students' behavior data; The information processing module is used to: decrypt the data through a trusted execution environment, and construct a multi-modal fusion analysis model to perform spatio-temporal feature extraction and cross-modal correlation analysis on the decrypted students' behavior data to obtain students' behavior feature vectors and behavior association rule sets; The behavior portrait module is used to: generate a dynamically updated student behavior portrait according to the students' behavior feature vectors and the behavior association rule sets, and in combination with an adaptive clustering model based on an improved K-means algorithm, and the student behavior portrait includes quantization labels of learning concentration, activity participation degree, and consumption tendency; The visualization module is used to: output the student behavior portrait to a management terminal through an encrypted API interface, and implement hierarchical access control based on a role permission matrix, and the role permission matrix defines the access levels of three types of roles, namely administrators, teachers, and counselors, to the portrait data.

[0005] Furthermore, the multi-source sensor includes an infrared behavior capture sensor, a classroom panoramic camera, an RFID positioning device, and a campus one-card terminal, and the mobile terminal device includes a smart bracelet and a campus APP client.

[0006] Furthermore, the information collection module is further configured to: Perform data cleaning, noise suppression, and anomaly behavior detection based on a sliding time window on the student behavior data. The anomaly behavior detection adopts an isolation forest algorithm and a dynamic threshold cooperation mechanism. When an anomaly is detected, the dynamic encryption policy of the secure transmission module is upgraded synchronously.

[0007] Furthermore, the dynamic encryption policy of the secure transmission module includes: Divide the student behavior data into data blocks of a fixed length, and independently encrypt each data block using the SM4 algorithm; generate a dynamic session key based on the SM2 algorithm, and complete end-to-end key distribution through the two-way authentication channel; set a decryption whitelist in the trusted execution environment, and only allow authorized analysis processes in the information processing module to access the decrypted student behavior data.

[0008] Furthermore, the multi-modal fusion analysis model includes: a time series analysis unit, a cross-modal association unit, a feature fusion unit, and a rule generation unit; among them, the time series analysis unit is used to: extract time-dependent features in the decrypted student behavior data based on a bidirectional LSTM neural network; the cross-modal association unit is used to: based on the Transformer module of the multi-head attention mechanism, input the cross-modal decrypted student behavior data, including learning activity data, classroom behavior data, and consumption trajectory data, and generate a cross-modal association feature vector by calculating cross-modal attention weights; the feature fusion unit is used to: perform weighted splicing on the time-dependent features and the cross-modal association feature vector to generate the student behavior feature vector; the rule generation unit is used to: adopt Granger causality test to mine behavior trigger rules from the student behavior feature vector, and form the behavior association rule set including early warnings of decreased learning efficiency, abnormal consumption, and sudden changes in classroom participation.

[0009] Further, the adaptive clustering model includes: an initial clustering center optimization unit, a dynamic weight adjustment unit, and a portrait update trigger unit; wherein, the initial clustering center optimization unit is used to: calculate the candidate center weights by fusing the silhouette coefficient and the density peak value, and select the top K data points with the highest weight values from the student behavior feature vectors as the initial centers; the dynamic weight adjustment unit is used to: generate a weight factor γ = α × β according to the time decay coefficient α and the category discrimination degree β of the student behavior data, and is used for the update calculation of the clustering center during the iteration process; the portrait update trigger unit is used to: when the cosine similarity between the student behavior feature vector and the existing clustering center is lower than a preset threshold, start the portrait dynamic update process.

[0010] Further, the hierarchical access control includes: defining that the administrator can access all quantization labels, the teacher can only access the desensitized quantization labels of learning concentration and activity participation, and the counselor can only access the desensitized quantization labels of activity participation.

[0011] Further, the learning activity data includes library entry and exit records, online learning duration, and homework submission time; the classroom behavior data includes the head-up rate, interactive response frequency, and position movement trajectory; the consumption trajectory data includes food consumption amount, supermarket consumption frequency, and recharge time distribution.

[0012] Further, the visualization module is specifically used to: Divide the quantization labels of personal identity identification and consumption tendency into identity-related categories, apply a combined strategy of field replacement and differential privacy perturbation to the identity-related category labels, and divide the quantization labels of learning concentration and activity participation into behavior feature categories, and apply a numerical generalization strategy to the behavior feature category labels.

[0013] Further, it also includes: a data traceability module; The data traceability module is used to: conduct blockchain evidence storage on the collection, encrypted transmission, processing, and portrait generation processes of student behavior data, and provide an audit traceability interface based on the hash value in the visualization interface module.

[0014] Compared with the prior art, the student information management system of the present invention has the following beneficial effects: The system of the present invention realizes the full-dimensional real-time collection and secure interaction of students' behavior data through multi-source sensor fusion and dynamic encryption transmission technology, breaking through the limitations of traditional systems in the management of single data types; based on multi-modal fusion analysis and a trusted computing environment, it constructs the ability to mine spatio-temporal features and cross-scene associations of students' behavior, improving the intelligence level of data processing; through an adaptive behavior portrait and hierarchical access mechanism, it generates quantifiable behavior tags that can be dynamically updated, providing precise decision-making support for education administrators while ensuring the secure and controllable use of sensitive information.

[0015] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be learned from the practice of the present invention. Brief Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 It is a schematic structural diagram of a student information management system. Detailed Embodiments

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0019] Figure 1 It shows a schematic structural diagram of an embodiment of a student information management system 100 provided by the present invention. As Figure 1 shown, the system 100 includes: an information collection module 110, a secure transmission module 120, an information processing module 130, a behavior portrait module 140, and a visualization module 150; The information collection module 110 is used to: collect students' behavior data in real time through multi-source sensors and mobile terminal devices deployed in the campus scenario, and the students' behavior data includes personal identity identifiers, learning activity data, classroom behavior data, and consumption track data.

[0020] Among them, the personal identity identifier is defaulted to the student number, and the learning activity data includes library entry and exit records, online learning duration, and homework submission time; classroom behavior data includes the head-up rate, interactive response frequency, and position movement trajectory; consumption trajectory data includes food consumption amount, supermarket consumption frequency, and recharge time distribution.

[0021] Among them, the multi-source sensors include infrared behavior capture sensors, classroom panoramic cameras, RFID positioning devices, and campus one-card terminals, and the mobile terminal devices include smart bracelets and campus APP clients. The overall layout principle of the multi-source sensors and mobile terminal devices is as follows: ① Cover all scenarios of students' learning, life, and consumption to ensure multi-dimensional collection of behavior data. ② Avoid collecting data in sensitive areas (such as inside the dormitory and the bathroom), and anonymize key data. ③ Based on the campus 5G / Wi-Fi 6 network, achieve low-latency data transmission and device linkage. Taking the classroom scenario as an example, one group of infrared behavior capture sensors can be installed on each of the front and rear walls of the classroom, covering all the seats in the classroom, and used to monitor the head-up rate, seat leaving frequency, limb movements, etc. of students. For example, if it is detected that the head-up rate of student A is lower than 40% during a 45-minute class, it is marked as distracted. The classroom panoramic camera is set on the ceiling in the center of the classroom, and the wide-angle lens covers a 360° field of view, and is used to record the interactive response frequency of students and the participation degree of group discussions. The picture is processed in real time by the edge computing node to be blurred, and only the action outline is retained. The RFID positioning device is installed above the classroom door frame and is bound to the campus one-card, and is used to record the entry and exit time of students, and analyze the position preference in combination with the seat coordinates (such as the second column in the first row). The smart bracelet is used to monitor the heart rate change (tension / relaxation state) and the seat movement trajectory (such as updating the coordinates every 5 seconds). For example, if the heart rate of student B is higher than 100 beats per minute for 20 consecutive minutes, then an association analysis is performed on student B to determine whether it is exam anxiety. Taking the cafeteria and consumption scenario as an example, the campus APP client is installed at the cafeteria window, supermarket cashier desk, and vending machine, and is used to record the consumption amount, time, and category (such as having breakfast for 8 yuan, including soy milk and steamed buns). When the smart bracelet is close to the POS machine to complete the payment, it synchronously records the consumption location and time.

[0022] It should be noted that the campus scenarios include but are not limited to: classroom scenarios, library scenarios, cafeteria scenarios, sports scenarios, and dormitory scenarios. The clocks of all device equipment are synchronized to the campus NTP server, with an error of <1 millisecond. For example, the classroom behavior of student C (the head-down rate increases at 11:00) is associated with the cafeteria consumption data (not dining at 11:30), and it is inferred that there is an abnormal health condition. Another example is that when the RFID positioning device detects that a certain student has not left the dormitory for 24 hours, it links the smart electricity meter data (the power consumption approaches zero) to trigger the counselor to check. To protect the privacy and security of students, in this embodiment, the personal identity identifier is replaced with an anonymous hash value (such as student number 20231001 is replaced with: 7a3b9c), and only the action skeleton of the camera picture is retained, and the facial features are deleted.

[0023] In an alternative approach, the information collection module 110 is further configured to: Perform data cleaning, noise suppression, and anomaly behavior detection based on a sliding time window on the student behavior data. The anomaly behavior detection adopts an isolation forest algorithm and a dynamic threshold collaboration mechanism. When an anomaly is detected, the dynamic encryption policy of the secure transmission module is triggered for upgrade synchronously.

[0024] Among them, the sliding time window refers to dividing the data stream according to a fixed time interval, such as 10 minutes. The isolation forest algorithm is an anomaly detection algorithm based on a tree structure that identifies data points deviating from the mainstream distribution. The ways of anomaly detection include: spatio-temporal anomaly (such as library access records at 3 am), frequency anomaly (such as the campus card being swiped continuously 10 times within 1 minute), association anomaly (such as a 90% head-down rate in class but accompanied by a high frequency of interactive responses), and physiological anomaly (such as the intelligent bracelet detecting a continuous high heart rate but no exercise record).

[0025] The secure transmission module 120 is configured to: establish a two-way authentication channel between the information collection module 110 and the information processing module 130, and perform segmented encryption on the student behavior data by using a dynamic encryption policy based on the national cryptographic algorithm to obtain the encrypted student behavior data.

[0026] In an alternative approach, the dynamic encryption policy of the secure transmission module 120 includes: Divide the student behavior data into data blocks of a fixed length, and independently encrypt each data block by using the SM4 algorithm; generate a dynamic session key based on the SM2 algorithm, and complete end-to-end key distribution through the two-way authentication channel; set a decryption whitelist in the trusted execution environment, and only allow authorized analysis processes in the information processing module to access the decrypted student behavior data.

[0027] Among them, the SM4 algorithm is a national cryptographic symmetric encryption algorithm for data block encryption. The decryption whitelist refers to a security policy that only allows authorized processes to access decrypted data.

[0028] The information processing module 130 is configured to: decrypt the data through the trusted execution environment, and construct a multi-modal fusion analysis model to perform spatio-temporal feature extraction and cross-modal association analysis on the decrypted student behavior data to obtain student behavior feature vectors and behavior association rule sets.

[0029] Among them, the multi-modal fusion analysis model is a machine learning model that integrates different data modalities (such as time series, behavior associations) for extracting composite features. The student behavior feature vector is a structured numerical set that extracts the key features of student behavior through multi-modal data analysis and is used to quantitatively describe their comprehensive performance in learning, consumption, classroom participation, etc. This vector is usually stored in the form of an array with a fixed dimension for easy processing by machine learning models. The student behavior feature vector includes: learning behavior features, classroom behavior features, consumption behavior features, and cross-modal features. The behavior association rule set is a set of causal or strongly associated relationships that mines the potential connections between student behavior patterns through statistical analysis (such as Granger causality test) and is used to predict risks or optimize management strategies. Each rule contains a precondition, a conclusion, and a confidence level.

[0030] In an optional manner, the multi-modal fusion analysis model includes: a time series analysis unit, a cross-modal association unit, a feature fusion unit, and a rule generation unit. The time series analysis unit is used to: extract the time-dependent features in the decrypted student behavior data based on a bidirectional LSTM neural network; the cross-modal association unit is used to: based on the Transformer module with a multi-head attention mechanism, input the decrypted student behavior data across modalities, including learning activity data, classroom behavior data, and consumption trajectory data, and generate a cross-modal association feature vector by calculating the cross-modal attention weights; the feature fusion unit is used to: perform weighted splicing on the time-dependent features and the cross-modal association feature vector to generate the student behavior feature vector; the rule generation unit is used to: adopt the Granger causality test to mine the behavior trigger rules from the student behavior feature vector and form the behavior association rule set including early warnings for decreased learning efficiency, abnormal consumption, and sudden changes in classroom participation.

[0031] It should be noted that the bidirectional LSTM refers to a recurrent neural network that can capture the dependencies before and after in a time series. The Granger causality test is used to judge the causal relationship between variables, such as whether homework delay leads to a decrease in grades. The multi-modal fusion analysis model extracts time features (such as the trend of learning duration) through LSTM, and the Transformer calculates cross-modal associations (such as the relationship between classroom behavior and consumption), and generates a feature vector and warning rules after fusion. For example, it is analyzed that there is a causal relationship between a student's late submission of homework (learning activity data) and a decrease in classroom participation 3 days later (classroom behavior data), and a warning rule is generated.

[0032] The behavior portrait module 140 is used to: generate a dynamically updated student behavior portrait according to the student behavior feature vector and the behavior association rule set, and in combination with an adaptive clustering model based on an improved K-means algorithm. The student behavior portrait includes quantitative labels for learning concentration, activity participation, and consumption tendency.

[0033] In an alternative approach, the adaptive clustering model includes: an initial clustering center optimization unit, a dynamic weight adjustment unit, and a portrait update trigger unit. Among them, the initial clustering center optimization unit is used to: calculate the candidate center weights by fusing the silhouette coefficient and the density peak, and select the top K data points with the highest weight values from the student behavior feature vectors as the initial centers; the dynamic weight adjustment unit is used to: generate a weight factor γ = α × β according to the time decay coefficient α and the class discrimination degree β of the student behavior data, and use it for the update calculation of the clustering center during the iteration process; the portrait update trigger unit is used to: when the cosine similarity between the student behavior feature vector and the existing clustering center is lower than a preset threshold, start the portrait dynamic update process.

[0034] Specifically: ① Perform Z-score standardization on the student behavior feature vectors to eliminate the dimension difference. ② Calculate the density peak (average distance to neighboring points) and silhouette coefficient (clustering tightness) of each data point, and select the data points with density peak > threshold and silhouette coefficient > 0.5 as candidate centers. ③ Sort the candidate centers in descending order of the weight value (weight = density peak × silhouette coefficient), and select the top K points as the initial centers. ④ Define the time decay coefficient α of the data weight over time, and calculate the Mahalanobis distance between the current clustering center and the global mean, β = distance / global standard deviation; during the iteration process, dynamically adjust the influence weight of the data points on the clustering center according to γ. ⑤ Convert the high-confidence (> 80%) association rules into clustering constraint conditions, and perform manual intervention or automatic adjustment on the clustering results that violate the association rules. ⑥ Calculate the cosine similarity between the new data points and the existing clustering centers; if the similarity < threshold (such as 0.5) continuously exceeds a preset period (such as 3 days), trigger the portrait update; each clustering center corresponds to a type of behavior portrait, and generate readable labels according to the eigenvalue of the clustering center; synchronize the newly collected data every 30 minutes and incrementally update the clustering model.

[0035] The visualization module 150 is used to: output the student behavior portrait to the management terminal through an encrypted API interface, and implement hierarchical access control based on the role permission matrix, where the role permission matrix defines the access levels of three types of roles, namely administrators, teachers, and counselors, to the portrait data.

[0036] Among them, the hierarchical access control includes: defining that administrators can access all quantitative labels, teachers can only access the desensitized quantitative labels of learning concentration and activity participation, and counselors can only access the desensitized quantitative labels of activity participation.

[0037] In an alternative approach, the visualization module 150 is specifically used to: Divide the quantitative tags of personal identity identification and consumption tendency into identity - related categories, apply a combined strategy of field replacement and differential privacy perturbation to the identity - related category tags, and divide the quantitative tags of learning concentration and activity participation into behavior - characteristic categories, and apply a numerical generalization strategy to the behavior - characteristic category tags.

[0038] Among them, personal identity identification specifically includes: student ID, name, and campus card ID. Consumption - tendency quantitative tags include: consumption amount, consumption time, and consumption location. Learning - concentration quantitative tags include: classroom head - up rate, and homework - completion efficiency; Activity - participation quantitative tags include: club - activity attendance rate, and extracurricular - practice times. The desensitization strategy for identity - related category tags includes field replacement and differential privacy perturbation. The desensitization strategy for behavior - characteristic category tags includes numerical generalization processing. Through the above steps, while protecting students' privacy, it ensures the availability and compliance of data in educational management.

[0039] In an optional manner, it further includes: a data traceability module; The data traceability module is used for: conducting blockchain evidence - storage on the process of collecting, encrypting and transmitting, processing, and generating portraits of students' behavior data, and providing an audit - trace interface based on hash values in the visualization interface module.

[0040] Among them, in the data - collection stage, sensor / terminal device ID, collection timestamp, data type, etc. are recorded. In the encrypted - transmission stage, encryption algorithm (SM4), key version, transmission start - and - end times are recorded. In the processing stage, TEE decryption operation, analysis - model version, and feature - vector generation time are recorded. In the portrait - generation stage, clustering - algorithm version, portrait tags, and update time are recorded.

[0041] The technical solution of this embodiment realizes the full - dimension real - time collection and secure interaction of students' behavior data through multi - source sensor fusion and dynamic encryption - transmission technology, breaking through the limitations of traditional systems in the management of single - data types; constructs the spatio - temporal feature mining and cross - scenario association capabilities of students' behavior based on multi - modal fusion analysis and trusted computing environment, improving the intelligent level of data processing; generates quantifiable behavior tags that can be dynamically updated through adaptive behavior portraits and hierarchical access mechanisms, providing precise decision - making support for educational administrators, and at the same time ensuring the secure and controllable use of sensitive information.

[0042] A student information management method provided by the present invention includes the following steps: S1. Real - time collect students' behavior data through multi - source sensors and mobile terminal devices deployed in campus scenarios, where the students' behavior data includes personal identity identification, learning - activity data, classroom - behavior data, and consumption - trajectory data; S2. Establish a two-way authentication channel, and segmentally encrypt the student behavior data using a dynamic encryption strategy based on national cryptographic algorithms to obtain the encrypted student behavior data; S3. Decrypt the data through a trusted execution environment, and construct a multi-modal fusion analysis model to extract spatio-temporal features and perform cross-modal correlation analysis on the decrypted student behavior data to obtain student behavior feature vectors and a set of behavior association rules; S4. According to the student behavior feature vectors and the set of behavior association rules, and in combination with an adaptive clustering model based on an improved K-means algorithm, generate a dynamically updated student behavior portrait, where the student behavior portrait includes quantization labels for learning concentration, activity participation, and consumption tendency; S5. Output the student behavior portrait to a management terminal through an encrypted API interface, and implement hierarchical access control based on a role permission matrix, where the role permission matrix defines the access levels of three types of roles, namely administrators, teachers, and counselors, to the portrait data.

[0043] In an alternative embodiment, the multi-source sensors include infrared behavior capture sensors, classroom panoramic cameras, RFID positioning devices, and campus one-card terminals, and the mobile terminal devices include smart bracelets and campus APP clients.

[0044] In an alternative embodiment, it further includes: Perform data cleaning, noise suppression, and anomaly behavior detection based on a sliding time window on the student behavior data. The anomaly behavior detection uses an isolation forest algorithm and a dynamic threshold cooperation mechanism, and when an anomaly is detected, the dynamic encryption strategy of the security transmission module is synchronously upgraded.

[0045] In an alternative embodiment, the dynamic encryption strategy includes: Divide the student behavior data into data blocks of a fixed length, and independently encrypt each data block using the SM4 algorithm; generate a dynamic session key based on the SM2 algorithm, and complete end-to-end key distribution through the two-way authentication channel; set a decryption whitelist in the trusted execution environment, and only allow authorized analysis processes in the information processing module to access the decrypted student behavior data.

[0046] In an alternative approach, the multimodal fusion analysis model includes: a time series analysis unit, a cross-modal association unit, a feature fusion unit, and a rule generation unit; wherein, the time series analysis unit is configured to: extract time-dependent features from the decrypted student behavior data based on a bidirectional LSTM neural network; the cross-modal association unit is configured to: input the decrypted student behavior data across modalities, including learning activity data, classroom behavior data, and consumption trajectory data, and generate a cross-modal association feature vector by calculating cross-modal attention weights based on a Transformer module with a multi-head attention mechanism; the feature fusion unit is configured to: perform weighted concatenation on the time-dependent features and the cross-modal association feature vector to generate the student behavior feature vector; the rule generation unit is configured to: use Granger causality test to mine behavior trigger rules from the student behavior feature vector and form the behavior association rule set including early warnings of decreased learning efficiency, abnormal consumption, and sudden changes in classroom participation.

[0047] In an alternative approach, the adaptive clustering model includes: an initial clustering center optimization unit, a dynamic weight adjustment unit, and a portrait update trigger unit; wherein, the initial clustering center optimization unit is configured to: calculate candidate center weights by fusing the silhouette coefficient and density peaks, and select the top K data points with the highest weight values from the student behavior feature vector as the initial centers; the dynamic weight adjustment unit is configured to: generate a weight factor γ = α × β according to the time decay coefficient α and the category discrimination degree β of the student behavior data for the update calculation of the clustering centers during the iteration process; the portrait update trigger unit is configured to: initiate the portrait dynamic update process when the cosine similarity between the student behavior feature vector and the existing clustering centers is lower than a preset threshold.

[0048] In an alternative approach, the hierarchical access control includes: defining that administrators can access all quantization labels, teachers can only access the quantization labels of the desensitized learning concentration and activity participation, and counselors can only access the quantization labels of the desensitized activity participation.

[0049] In an alternative approach, the learning activity data includes library entry and exit records, online learning duration, and homework submission time; the classroom behavior data includes the head-up rate, interactive response frequency, and location movement trajectory; the consumption trajectory data includes the catering consumption amount, supermarket consumption frequency, and recharge time distribution.

[0050] In an alternative approach, it further includes: Divide the quantification tags of personal identity identifiers and consumption preferences into identity-related categories, apply a combined strategy of field replacement and differential privacy perturbation to the identity-related category tags, and divide the quantification tags of learning focus and activity participation into behavioral feature categories, and apply a numerical generalization strategy to the behavioral feature category tags.

[0051] In an alternative approach, it further includes: To conduct blockchain-based evidence preservation for the processes of collecting, encrypting and transmitting, processing, and generating portraits of student behavior data, and provide an audit traceability interface based on hash values in the visualization interface module.

[0052] The above description is only the preferred embodiment of the present invention and an explanation of the applied technical principles. Those skilled in the art should understand that the disclosed scope in the present invention is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the technical solutions formed by mutually replacing the above features with the (but not limited to) technical features with similar functions disclosed in the present invention.

[0053] It should be noted that the terms "first", "second", etc. in the description and claims of this application are used to distinguish similar objects, and represent a limitation on a specific order or sequence. Under appropriate circumstances, the order of use of similar objects can be interchanged so that the embodiments of this application described here can be implemented in an order other than the illustrated or described order.

[0054] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A student information management system, characterized in that, Including: An information collection module, a secure transmission module, an information processing module, a behavior profiling module, and a visualization module; The information collection module is used to: collect students' behavior data in real time through multi-source sensors and mobile terminal devices deployed in the campus scenario, and the students' behavior data includes personal identity identifiers, learning activity data, classroom behavior data, and consumption trajectory data; The secure transmission module is used to: establish a two-way authentication channel between the information collection module and the information processing module, and segmentally encrypt the students' behavior data using a dynamic encryption strategy based on national cryptographic algorithms to obtain encrypted students' behavior data; The information processing module is used to: decrypt the data through a trusted execution environment, and construct a multi-modal fusion analysis model to extract spatio-temporal features and perform cross-modal correlation analysis on the decrypted students' behavior data to obtain students' behavior feature vectors and behavior association rule sets; The behavior profiling module is used to: generate a dynamically updated students' behavior profile according to the students' behavior feature vectors and the behavior association rule sets, and in combination with an adaptive clustering model based on an improved K-means algorithm, and the students' behavior profile includes quantization labels of learning concentration, activity participation, and consumption tendency; The visualization module is used to: output the students' behavior profile to the management terminal through an encrypted API interface, and implement hierarchical access control based on a role permission matrix, and the role permission matrix defines the access levels of three types of roles, namely administrators, teachers, and counselors, to the profile data.

2. The student information management system according to claim 1, characterized in that The multi-source sensors include infrared behavior capture sensors, classroom panoramic cameras, RFID positioning devices, and campus one-card terminals, and the mobile terminal devices include smart bracelets and campus APP clients.

3. The student information management system according to claim 1, characterized in that The information collection module is further used to: Perform data cleaning, noise suppression, and anomaly behavior detection based on a sliding time window on the students' behavior data, and the anomaly behavior detection adopts an isolation forest algorithm and a dynamic threshold cooperation mechanism, and when an anomaly is detected, the dynamic encryption strategy upgrade of the secure transmission module is synchronously triggered.

4. The student information management system according to claim 1, characterized in that, The dynamic encryption strategy of the secure transmission module includes: Dividing the students' behavior data into data blocks of a fixed length, independently encrypting each data block using the SM4 algorithm; generating a dynamic session key based on the SM2 algorithm, and completing end-to-end key distribution through the two-way authentication channel; setting a decryption whitelist in the trusted execution environment, and only allowing authorized analysis processes in the information processing module to access the decrypted students' behavior data.

5. The student information management system according to claim 1, wherein The multimodal fusion analysis model includes: a time series analysis unit, a cross-modal association unit, a feature fusion unit, and a rule generation unit; among them, the time series analysis unit is used to: extract time-dependent features in the decrypted student behavior data based on a bidirectional LSTM neural network; the cross-modal association unit is used to: based on the Transformer module of the multi-head attention mechanism, input the cross-modal decrypted student behavior data, including learning activity data, classroom behavior data, and consumption trajectory data, and generate a cross-modal association feature vector by calculating cross-modal attention weights; the feature fusion unit is used to: perform weighted splicing on the time-dependent features and the cross-modal association feature vector to generate the student behavior feature vector; the rule generation unit is used to: adopt Granger causality test to mine behavior trigger rules from the student behavior feature vector, and form the behavior association rule set including early warnings of decreased learning efficiency, abnormal consumption warnings, and sudden changes in classroom participation warnings.

6. The student information management system according to claim 1, characterized in that, The adaptive clustering model includes: an initial clustering center optimization unit, a dynamic weight adjustment unit, and a portrait update trigger unit; among them, the initial clustering center optimization unit is used to: calculate candidate center weights by fusing the silhouette coefficient and density peak value, and select the top K data points with the highest weight values from the student behavior feature vector as the initial centers; the dynamic weight adjustment unit is used to: generate a weight factor γ = α × β according to the time decay coefficient α and category discrimination degree β of the student behavior data for the update calculation of the clustering center during the iteration process; the portrait update trigger unit is used to: when the cosine similarity between the student behavior feature vector and the existing clustering center is lower than a preset threshold, start the portrait dynamic update process.

7. The student information management system according to claim 1, wherein, The hierarchical access control includes: defining that administrators can access all quantization labels, teachers can only access the quantization labels of desensitized learning concentration and activity participation, and counselors can only access the quantization labels of desensitized activity participation.

8. The student information management system according to claim 1, characterized in that, The learning activity data includes library access records, online learning duration, and homework submission time; the classroom behavior data includes head-up rate, interactive response frequency, and position movement trajectory; the consumption trajectory data includes food consumption amount, supermarket consumption frequency, and recharge time distribution.

9. The student information management system according to claim 7, characterized in that The visualization module is specifically used for: Dividing the quantization labels of personal identity identification and consumption tendency into identity association categories, applying a combined strategy of field replacement and differential privacy perturbation to the identity association category labels, and dividing the quantization labels of learning concentration and activity participation into behavior feature categories, and applying a numerical generalization strategy to the behavior feature category labels.

10. The student information management system according to any one of claims 1-9, characterized in that, It also includes: A data traceability module; The data traceability module is used to: conduct blockchain evidence preservation on the collection, encrypted transmission, processing, and portrait generation processes of student behavior data, and provide an audit traceability interface based on hash values in the visualization interface module.

Images