Student mental health digital intelligence early warning method

Abstract

The invention discloses a student psychological health digital intelligence early warning method, which belongs to the technical field of big data, and comprises the steps of establishing a data acquisition layer, a data preprocessing layer, a data analysis and modeling layer and a psychological crisis early warning layer in a distributed server cluster. According to the invention, a more comprehensive psychological health assessment basis is provided, real-time updating and processing of data can be realized, timeliness and accuracy of data analysis are improved, psychological health states of students can be assessed more comprehensively, abnormal behaviors of the students can be accurately identified, and therefore, the students can have a better psychological health assessment effect. And according to the frequency and the grade of the abnormal behavior, adjusting the influence of the abnormal behavior on the psychological health score.

Description

Technical Field

[0001] The present invention belongs to the field of big data technology, and in particular relates to a digital intelligence early warning method for students' mental health. Background Art

[0002] In today's society, students' mental health issues are receiving increasing attention, especially in campus environments, where mental health issues may have a significant impact on students' academic performance, quality of life, and overall development. In order to effectively monitor and warn students' mental health, existing technologies mainly rely on traditional questionnaires, psychological tests, and manual analysis. However, these traditional methods have the following disadvantages:

[0003] Data collection limitations: Traditional mental health assessments usually rely on paper questionnaires or regular psychological tests. These methods cannot collect students’ mental state data in real time, nor can they capture students’ dynamic changes in their daily lives.

[0004] Long evaluation cycle: The existing psychological crisis warning and intervention systems in colleges and universities mostly use traditional psychological scale monitoring data, and only make passive reactive interventions after the psychological crisis becomes prominent, which has problems such as subjectivity and lag;

[0005] Strong subjectivity: Traditional questionnaires and tests are highly influenced by the subjective responses of respondents, which may lead to inaccuracy and consistency issues in the results;

[0006] Difficulty in data integration: Existing methods often treat mental health data separately from other relevant data (such as academic performance, behavioral records, etc.), making it difficult to comprehensively consider the impact of multiple factors on mental health; in the traditional crisis reporting process, relevant personnel lack common sense about mental health and lack of sensitivity, which leads to underreporting and false reporting.

[0007] Single data source or specific psychological screening method: In the existing mental health early warning system, it usually relies only on a single data source or specific psychological screening method, which can easily lead to an incomplete assessment of the students' mental health status, prone to deviations, and unable to identify potential psychological crises in a timely and accurate manner. Summary of the invention

[0008] The purpose of this invention is to provide a digital early warning method for students' mental health, which solves the technical problems of establishing a multi-dimensional and multi-level mental health assessment model and improving the accuracy of mental health assessment.

[0009] To achieve the above object, the present invention adopts the following technical solution:

[0010] A digital intelligence early warning method for students' mental health includes the following steps:

[0011] Step 1: Establish a digital early warning platform for students' mental health, including a distributed server cluster, in which a data collection layer, a data preprocessing layer, a data analysis and modeling layer, and a psychological crisis early warning layer are established;

[0012] Step 2: The data collection layer collects students’ mental health data from multiple data sources. Mental health data includes behavioral data and psychological data.

[0013] Step 3: The data preprocessing layer preprocesses the mental health data;

[0014] Integrate all pre-processed mental health data into a unified database to provide support for subsequent data analysis and model training;

[0015] Step 4: The data analysis and modeling layer extracts features from each type of mental health data, and establishes a scoring model for each type of data. The weight of the scores is calculated based on all the scoring models, and an early warning model for psychological crisis prediction is constructed.

[0016] Step 5: Use the data in the database to train the early warning model;

[0017] Step 6: Use the early warning model to evaluate the student's mental state, generate an evaluation score, and generate early warning information based on the score.

[0018] Preferably, when executing step 1, the multiple data sources include a campus card system, a mobile phone APP, an educational system, a campus gate access control system, and a dormitory management system;

[0019] Behavioral data include economic consumption, academic performance, subject attendance records, campus entry and exit records, dormitory return records, and dormitory behavior records;

[0020] Among them, economic consumption information is collected by the campus card system, academic performance and subject attendance records are collected by the academic affairs system, campus entry and exit behavior records are collected by the campus gate access control system, dormitory return records are collected by the dormitory management system's access control system, and dormitory behavior records are behavior records entered by dormitory managers through the dormitory management system;

[0021] Psychological data include the results of regular mental health screening, which are collected by mobile phone APP through questionnaires.

[0022] Preferably, when executing step 4, the early warning model for psychological crisis prediction is as follows:

[0023] P total =w consumption P consumption +w academic P academic+w behavior P behavior +w survey P survey ;

[0024] Among them, w consumption is the weight of the consumption model, w academic is the weight of the academic model, w behavior is the weight of the behavior model, w survey Give weight to regular screening results for mental health.

[0025] Preferably, the economic consumption situation includes the consumption amount x1 and the consumption frequency x2, and a consumption-psychological score list Sx is preset, and the standardized values ​​of the consumption amount x1 and the consumption frequency x2, i.e., z1 and z2, are calculated according to the following formula:

[0026]

[0027] Among them, μ x1 is the average consumption amount, μ x2 is the average consumption amount, σ x1 is the standard deviation of consumption frequency, σ x2 is the standard deviation of consumption frequency;

[0028] The scoring model of economic consumption, namely the consumption model, is shown in the following formula:

[0029] S=w x1 ×z1+w x2 ×z2;

[0030] S is the comprehensive score corresponding to the economic consumption situation. According to the comprehensive score S, the consumption-psychological score list Sx is searched to obtain the corresponding mental health status score P. consumption .

[0031] Preferably, the academic performance and the subject attendance record are combined to adopt the same scoring model, namely, the academic model, which is specifically shown in the following formula:

[0032] P academic =α C ×C+β D ×D+γ;

[0033] Among them, P behavior is the mental health status score corresponding to the academic performance and subject attendance record, C is the score of academic performance, D is the score of attendance rate obtained according to the subject attendance record, α C , β D and γ are regression coefficients, a C The impact on the scores of coping with academic performance, β DCorresponding to the impact of the attendance score, γ is a constant term.

[0034] Preferably, the campus entry and exit behavior records, the dormitory return records and the dormitory behavior records are combined to adopt the same scoring model, namely the behavior model, and the specific formula is as follows:

[0035] P behavior =α1×Behavior+β1×(Entry×Attendance)+γ1×F rare +δ1×F frequent +∈1;

[0036] Among them, Behavior is the score of normal dormitory behavior records, Entry is the score of campus entry and exit behavior records, Attendance is the score of dormitory return records, and F rare is the influence coefficient of occasional abnormal behavior, F frequent is the influence coefficient of frequent abnormal behavior; P behavior is the final score of the behavior model; α1 is the regression coefficient of normal dormitory behavior records; β1 is the interaction coefficient of campus entry and exit behavior records and dormitory return records; γ1 is the influence coefficient of occasional abnormal behavior; δ1 is the influence coefficient of frequent abnormal behavior; ∈1 is a constant term.

[0037] Preferably, when executing step 6, according to P total The level of the warning information is generated, which specifically includes the following steps:

[0038] Step 6-1: Set the threshold of the score range and divide the risk according to the threshold of the score range: low risk [0, T1], medium risk (T1, T2], high risk T2 less than P total ; T1 and T2 are the upper and lower thresholds of the score range, respectively;

[0039] Step 6-2: Generate corresponding push information based on the risk classification results in step 6-1.

[0040] The digital intelligence early warning method for student mental health described in the present invention solves the technical problem of establishing a multi-dimensional and multi-level mental health assessment model and improving the accuracy of mental health assessment. The present invention integrates data from the campus card system, mobile phone APP, academic system, campus gate access control system and dormitory management system, including economic consumption, academic performance, subject attendance records, campus entry and exit behavior records, dormitory return records and dormitory behavior records. This comprehensive data collection method can fully understand the life and behavior status of students, provide a more comprehensive basis for mental health assessment, and through the distributed processing of the data acquisition layer, preprocessing layer, analysis and modeling layer, it can realize real-time update and processing of data, improve the timeliness and accuracy of data analysis, establish an economic consumption model, academic model, behavior model and mental health screening model, and comprehensively consider the impact of multiple factors on students' mental health. This multi-model fusion method can more comprehensively evaluate the mental health status of students, mark and process abnormal behaviors through behavioral models, accurately identify students' abnormal behaviors, and adjust their impact on mental health scores according to the frequency and level of abnormal behaviors, thereby providing more accurate early warning information, using early warning models to evaluate students' mental states, generate specific early warning information, and provide real-time information push for mental health warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 is a flow chart of the present invention;

[0042] Figure 2 It is a data flow schematic diagram of the present invention;

[0043] Figure 3 Schematic diagram of the system architecture of the present invention. DETAILED DESCRIPTION

[0044] like Figure 1-Figure 3 A digital intelligence early warning method for students' mental health is shown, comprising the following steps:

[0045] Step 1: Establish a digital early warning platform for students' mental health, including a distributed server cluster, in which a data collection layer, a data preprocessing layer, a data analysis and modeling layer, and a psychological crisis early warning layer are established;

[0046] In this embodiment, the data collection layer is responsible for collecting mental health data from various data sources (campus cards, mobile phone APPs, academic affairs systems, etc.), specifically including a data collection server, which communicates with various data sources through an API interface.

[0047] The data preprocessing layer is responsible for processing the raw data, including standardization, cleaning and other operations to ensure the quality and consistency of the data. It specifically includes a data preprocessing server and a database server.

[0048] The data analysis and modeling layer is responsible for feature extraction, building a scoring model, and calculating the score weights to form the final early warning model, which specifically includes the model building server.

[0049] The psychological crisis warning layer is responsible for evaluating students' psychological state according to the warning model and generating corresponding warning information, specifically including the psychological state evaluation server.

[0050] Multiple data sources include campus card system, mobile APP, academic affairs system, campus gate access control system and dormitory management system;

[0051] Behavioral data include economic consumption, academic performance, subject attendance records, campus entry and exit records, dormitory return records, and dormitory behavior records;

[0052] Among them, economic consumption information is collected by the campus card system, academic performance and subject attendance records are collected by the academic affairs system, campus entry and exit behavior records are collected by the campus gate access control system, dormitory return records are collected by the dormitory management system's access control system, and dormitory behavior records are behavior records entered by dormitory managers through the dormitory management system;

[0053] Psychological data include the results of regular mental health screening. The results of regular mental health screening are collected by mobile phone APP through questionnaires. Mobile phone APP is used to publish questionnaires for regular mental health screening. The questionnaires are ultimately filled out by students, teachers, and staff. Mobile phone APP scores the rules of the questionnaire and the students' behavioral rules.

[0054] In this embodiment, the data collection layer selects the campus card consumption system, the academic affairs management system, the campus gate access control system, the dormitory management system and the mobile phone APP as important data sources, covering students' economic consumption, learning performance, entry and exit behavior, and dormitory status.

[0055] Step 2: The data collection layer collects students’ mental health data from multiple data sources. Mental health data includes behavioral data and psychological data.

[0056] Behavioral data include economic consumption, academic performance, subject attendance records, campus entry and exit records, dormitory return records and dormitory behavior records; psychological data include regular mental health screening results (in the form of scores).

[0057] Step 3: The data preprocessing layer preprocesses the mental health data;

[0058] Integrate all pre-processed mental health data into a unified database to provide support for subsequent data analysis and model training;

[0059] In this embodiment, preprocessing includes removing redundancy and invalid data, standardizing features, unifying databases, etc.

[0060] In this embodiment, the feature extraction of data is to extract the features of data such as economic consumption, academic performance, and behavioral records, which are used for modeling of various models.

[0061] Step 4: The data analysis and modeling layer extracts features from each type of mental health data, and establishes a scoring model for each type of data. The weight of the scores is calculated based on all the scoring models, and an early warning model for psychological crisis prediction is constructed.

[0062] The early warning model for psychological crisis prediction is as follows:

[0063] P total =w consumption P consumption +w academic P academic +w behavior P behavior +w survey P survey ;

[0064] Among them, w consumption is the weight of the consumption model, w academic is the weight of the academic model, w behavior is the weight of the behavior model, w survey Give weight to regular screening results for mental health.

[0065] In this embodiment, the early warning model for psychological crisis prediction evaluates the students' mental health status by weighted combination of data from multiple different sources. Each type of data (such as consumption behavior, academic performance, behavior records, and mental health screening results) corresponds to a scoring model. These scoring models reflect the students' mental health status in specific dimensions. The final mental health status score is the weighted sum of all these model scores.

[0066] P total It represents the comprehensive mental health score of students and predicts whether students are likely to have a psychological crisis. Each weight reflects the relative importance of different dimensions to the overall mental health status. The weight setting can be adjusted based on actual data or expert experience.

[0067] The present invention integrates multiple factors so that the model can more comprehensively reflect the mental health status of students and avoids the evaluation bias that may be caused by relying solely on a single data source.

[0068] The economic consumption situation includes the consumption amount x1 and the consumption frequency x2. A consumption-psychological score list Sx is preset, and the standardized values ​​of the consumption amount x1 and the consumption frequency x2, i.e., z1 and z2, are calculated according to the following formula:

[0069]

[0070] Among them, μ x1 is the average consumption amount, μ x2 is the average consumption amount, σ x1 is the standard deviation of consumption frequency, σ x2 is the standard deviation of consumption frequency;

[0071] The scoring model of economic consumption, namely the consumption model, is shown in the following formula:

[0072] S=w x1 ×z1+w x2 ×z2;

[0073] S is the comprehensive score corresponding to the economic consumption situation. According to the comprehensive score S, the consumption-psychological score list Sx is searched to obtain the corresponding mental health status score P. consumption .

[0074] In this embodiment, the principle of the consumption model is to standardize the students' consumption behavior data (consumption amount and consumption frequency), and then obtain a comprehensive score reflecting the impact of consumption behavior on mental health through weighted summation.

[0075] Consumption amount x1 and consumption frequency x2 are two main consumption behavior indicators. They are standardized by subtracting the mean and then dividing by the standard deviation to eliminate the impact of different dimensions.

[0076] w x1 and w x2 It is the weight corresponding to the consumption amount and consumption frequency, reflecting the contribution of each indicator to the comprehensive score.

[0077] Through standardization and weighted summation, the model of the present invention can process data with different characteristics and convert them into a unified score, which is convenient for subsequent comprehensive analysis.

[0078] The academic performance and the subject attendance record are combined using the same scoring model, namely the academic model, as shown in the following formula:

[0079] P academic =α C ×C+β D ×D+γ;

[0080] Among them, P behavioris the mental health status score corresponding to the academic performance and subject attendance record, C is the score of academic performance, D is the score of attendance rate obtained according to the subject attendance record, α C , β D and γ are regression coefficients, α C The impact on the scores of coping with academic performance, β D Corresponding to the impact of the attendance score, γ is a constant term.

[0081] In this embodiment, the academic model combines students' academic performance and subject attendance records to assess the impact of these academic-related factors on mental health. The model weights and sums the scores of academic performance and attendance rate, and adds a constant term to obtain a comprehensive score that reflects the impact of students' academic performance on their mental health.

[0082] P behavior Represents the mental health status score corresponding to academic performance and subject attendance records.

[0083] C reflects the student's performance in the course.

[0084] D reflects students' attendance.

[0085] α C Represents the weight of the impact of academic performance on mental health status.

[0086] β D Represents the impact weight of attendance rate on mental health status.

[0087] γ is used as the baseline value or correction term of the model to adjust the overall score.

[0088] By comprehensively considering academic performance and attendance rate, the model of the present invention can evaluate the impact of academic-related factors on mental health from multiple angles, rather than relying solely on a single indicator. By adjusting the regression coefficient, it can adapt to the characteristics of different student groups or individuals, thereby improving the flexibility and accuracy of the model.

[0089] The campus entry and exit behavior records, the dormitory return records and the dormitory behavior records are combined to adopt the same scoring model, namely the behavior model, and the specific formula is as follows:

[0090] P behavior =α1×Behavior+β1×(Entry×Attendance)+γ1×F rare +δ1×F frequent +∈1;

[0091] Among them, Behavior is the score of normal dormitory behavior records, Entry is the score of campus entry and exit behavior records, Attendance is the score of dormitory return records, and Frare is the influence coefficient of occasional abnormal behavior, F frequent is the influence coefficient of frequent abnormal behavior; P behavior is the final score of the behavior model; α1 is the regression coefficient of normal dormitory behavior records; β1 is the interaction coefficient of campus entry and exit behavior records and dormitory return records; γ1 is the influence coefficient of occasional abnormal behavior; δ1 is the influence coefficient of frequent abnormal behavior; ∈1 is a constant term.

[0092] The behavioral model comprehensively evaluates the impact of students' daily behavior on their mental health by combining normal behavior scores and abnormal behavior corrections. It not only takes into account the normal performance of daily behavior, but also corrects abnormal behaviors, especially giving higher weight to frequent abnormal behaviors.

[0093] P behavior It reflects the overall impact of students' daily behavior on their mental health.

[0094] Behavior represents the normal behavior of students in the dormitory.

[0095] Entry×Attendance represents the interaction effect score between the campus entry and exit behavior records and the dormitory return records, which represents the comprehensive performance of students' on-campus and off-campus activities and dormitory return behaviors.

[0096] F rare It means that occasional abnormal behavior has little impact on mental health, that is, the score is small.

[0097] F frequent It means that frequently occurring abnormal behaviors have a greater impact on mental health, that is, the score is high.

[0098] By comprehensively considering normal behavior, interactive behavior and abnormal behavior, the model of the present invention can more accurately evaluate the impact of students' daily behavior on their mental health, especially when abnormal behavior occurs frequently, and can provide early warning more timely and effectively.

[0099] In this embodiment, abnormal behavior extraction is to identify abnormal behaviors from dormitory behavior records, and to classify abnormal behaviors into two categories: "occasional" and "frequent" according to the frequency of occurrence of the behaviors.

[0100] In an application scenario of this embodiment,

[0101] The criteria for defining abnormal behavior are as follows:

[0102] Failure to return to bed on time: Records of returning to bed later than the prescribed time;

[0103] Frequent entry and exit: frequently entering and leaving the dormitory or campus within a short period of time;

[0104] Behavioral abnormalities: Behaviors that do not fit normal patterns (such as excessive activity at night).

[0105] The specific method of marking abnormal behaviors is to mark and classify each abnormal behavior. The marking type is:

[0106] Occasionally: The behavior occurs less frequently, such as once a month;

[0107] Frequent: The behavior occurs frequently, such as weekly or more often.

[0108] In this embodiment, each abnormal behavior is marked and classified, and different scores are set.

[0109] Step 5: Use the data in the database to train the early warning model;

[0110] Step 6: Use the early warning model to evaluate the student's mental state, generate an evaluation score, and generate early warning information based on the score.

[0111] When executing step 6, according to P total The level of the warning information is generated, which specifically includes the following steps:

[0112] Step 6-1: Set the threshold of the score range and divide the risk according to the threshold of the score range: low risk [0, T1], medium risk (T1, T2], high risk T2 less than P total ; T1 and T2 are the upper and lower thresholds of the score range, respectively;

[0113] Step 6-2: Generate corresponding push information based on the risk classification results in step 6-1.

[0114] The digital intelligence early warning method for student mental health described in the present invention solves the technical problem of establishing a multi-dimensional and multi-level mental health assessment model and improving the accuracy of mental health assessment. The present invention integrates data from the campus card system, mobile phone APP, academic system, campus gate access control system and dormitory management system, including economic consumption, academic performance, subject attendance records, campus entry and exit behavior records, dormitory return records and dormitory behavior records. This comprehensive data collection method can fully understand the life and behavior status of students, provide a more comprehensive basis for mental health assessment, and through the distributed processing of the data acquisition layer, preprocessing layer, analysis and modeling layer, it can realize real-time update and processing of data, improve the timeliness and accuracy of data analysis, establish an economic consumption model, academic model, behavior model and mental health screening model, and comprehensively consider the impact of multiple factors on students' mental health. This multi-model fusion method can more comprehensively evaluate the mental health status of students, mark and process abnormal behaviors through behavioral models, accurately identify students' abnormal behaviors, and adjust their impact on mental health scores according to the frequency and level of abnormal behaviors, thereby providing more accurate early warning information, using early warning models to evaluate students' mental states, generate specific early warning information, and provide real-time information push for mental health warnings.

Claims

1. A digital intelligence early warning method for students' mental health, characterized by: The steps include: Step 1: Establish a digital early warning platform for students' mental health, including a distributed server cluster, in which a data collection layer, a data preprocessing layer, a data analysis and modeling layer, and a psychological crisis early warning layer are established; Step 2: The data collection layer collects students’ mental health data from multiple data sources. Mental health data includes behavioral data and psychological data. Step 3: The data preprocessing layer preprocesses the mental health data; Integrate all pre-processed mental health data into a unified database to provide support for subsequent data analysis and model training; Step 4: The data analysis and modeling layer extracts features from each type of mental health data, and establishes a scoring model for each type of data. The weight of the scores is calculated based on all the scoring models, and an early warning model for psychological crisis prediction is constructed. Step 5: Use the data in the database to train the early warning model; Step 6: Use the early warning model to evaluate the student's mental state, generate an evaluation score, and generate early warning information based on the score.

2. A method for early warning of student mental health according to claim 1, characterized in that: When executing step 1, multiple data sources include the campus card system, mobile phone APP, academic affairs system, campus gate access control system and dormitory management system; Behavioral data include economic consumption, academic performance, subject attendance records, campus entry and exit records, dormitory return records, and dormitory behavior records; Among them, economic consumption information is collected by the campus card system, academic performance and subject attendance records are collected by the academic affairs system, campus entry and exit behavior records are collected by the campus gate access control system, dormitory return records are collected by the dormitory management system's access control system, and dormitory behavior records are behavior records entered by dormitory managers through the dormitory management system; Psychological data include the results of regular mental health screening, which are collected by mobile phone APP through questionnaires.

3. A method for early warning of student mental health according to claim 2, characterized in that: When executing step 4, the early warning model for psychological crisis prediction is as follows: P total =w consumption P consumption +w academic P academic +w behavior P behavior +w survey P survey ; Among them, w consumption is the weight of the consumption model, w academic is the weight of the academic model, w behavior is the weight of the behavior model, w survey Give weight to regular screening results for mental health.

4. A method for early warning of student mental health according to claim 3, characterized in that: The economic consumption situation includes the consumption amount x1 and the consumption frequency x2. A consumption-psychological score list Sx is preset, and the standardized values ​​of the consumption amount x1 and the consumption frequency x2, i.e., z1 and z2, are calculated according to the following formula: Among them, μ x1 is the average consumption amount, μ x2 is the average consumption amount, σ x1 is the standard deviation of consumption frequency, σ x2 is the standard deviation of consumption frequency; The scoring model of economic consumption, namely the consumption model, is shown in the following formula: S=w x1 ×z1+w x2 ×z2; S is the comprehensive score corresponding to the economic consumption situation. According to the comprehensive score S, the consumption-psychological score list Sx is searched to obtain the corresponding mental health status score P. consumption .

5. A method for early warning of student mental health according to claim 3, characterized in that: The academic performance and the subject attendance record are combined using the same scoring model, namely the academic model, as shown in the following formula: P academic =a C ×C+β D ×D+γ;; Among them, P behavior is the mental health status score corresponding to the academic performance and subject attendance record, C is the score of academic performance, D is the score of attendance rate obtained according to the subject attendance record, α C ,β D and γ are regression coefficients, α C The impact on the scores of coping with academic performance, β D Corresponding to the impact of the attendance score, γ is a constant term.

6. A method for early warning of student mental health according to claim 3, characterized in that: The campus entry and exit behavior records, the dormitory return records and the dormitory behavior records are combined to adopt the same scoring model, namely the behavior model, and the specific formula is as follows: P behavior =α1×Behavior+β1×(Entry×Attendance)+γ1×F rare +δ1×F frequent +∈1; Among them, Behavior is the score of normal dormitory behavior records, Entry is the score of campus entry and exit behavior records, Attendance is the score of dormitory return records, and F rare is the influence coefficient of occasional abnormal behavior, F frequent is the influence coefficient of frequent abnormal behavior; P behavior is the final score of the behavior model; α1 is the regression coefficient of normal dormitory behavior records; β1 is the interaction coefficient of campus entry and exit behavior records and dormitory return records; γ1 is the influence coefficient of occasional abnormal behavior; δ1 is the influence coefficient of frequent abnormal behavior; ∈1 is a constant term.

7. A method for early warning of student mental health according to claim 3, characterized in that: When executing step 6, according to P total The level of the warning information is generated, which specifically includes the following steps: Step 6-1: Set the threshold of the score range and divide the risk according to the threshold of the score range: low risk [0, T1], medium risk (T1, T2], high risk T2 less than P total ; T1 and T2 are the upper and lower thresholds of the score range, respectively; Step 6-2: Generate corresponding push information based on the risk classification results in step 6-1.

Images