Enterprise digital risk prevention and control system construction system for mass health industry
By building a risk characteristic matrix and risk model, combining offline and real-time data, the problem of the contradiction between historical data and real-time data in the big health industry is solved, and the accuracy of risk prevention and control is improved.
Patent Information
- Application Number
- CN202510824324.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-19
- Publication Date
- 2025-08-19
AI Technical Summary
In the risk prevention and control of the big health industry, the timeliness contradiction between historical data and real-time data leads to a decrease in the accuracy of risk prevention and control.
Build a risk feature matrix to establish a module, collect offline historical features and real-time features through task scheduling and real-time monitoring, and build a risk feature matrix across time scales; establish offline and real-time risk models, integrate scores to output risk warning data; dynamically update data monitoring points.
The accuracy of risk prevention and control has been improved, and by combining historical and real-time data and dynamically adjusting monitoring points, the timeliness and accuracy of risk warnings has been enhanced.
Smart Images
Figure CN120508991A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of risk prevention and control technology, and in particular to a system for building a digital risk prevention and control system for enterprises in the health industry. Background Art
[0002] The big health industry encompasses multiple sectors, including healthcare, health care, elderly care, and insurance. Its operations involve core aspects such as user health data, medical service processes, and insurance product pricing. As the industry's digital transformation accelerates, risks such as data leaks, medical malpractice, and insurance fraud are becoming more frequent.
[0003] Currently, risk prevention and control requires monitoring of relevant data, including historical data and real-time data. During the risk prevention and control process, offline historical data such as historical transaction records and equipment maintenance logs can reflect long-term risk trends but cannot capture real-time anomalies. Real-time data such as sensor stream data and user behavior logs can perceive risks in a timely manner but lack the support of historical patterns. In summary, there is a contradiction in the timeliness between historical data and real-time data, which reduces the accuracy of risk prevention and control. Summary of the Invention
[0004] The purpose of the present invention is to provide a digital risk prevention and control system for enterprises in the big health industry, aiming to solve the technical problem in the existing technology that in the process of risk prevention and control, there is a contradiction in the timeliness between historical data and real-time data, thereby reducing the accuracy of risk prevention and control.
[0005] To achieve the above objectives, the present invention adopts a system for building a digital risk prevention and control system for enterprises in the health industry, including a risk feature matrix establishment module, a risk early warning module, and a monitoring point dynamic update module; wherein:
[0006] The risk feature matrix establishment module is used to trigger offline data collection and real-time data collection requests, obtain offline historical features and real-time features, and construct a risk feature matrix across time scales;
[0007] The risk warning module is used to establish an offline risk model and a real-time risk model, and output risk warning data for the data collection business ID;
[0008] The monitoring point dynamic update module is used to obtain risk warning data and dynamically update data monitoring points.
[0009] The risk feature matrix establishment module includes an offline data acquisition module, a real-time data acquisition module, and a feature alignment module; wherein:
[0010] The offline data collection module is used to use task scheduling to regularly trigger offline data collection tasks, generate a unique identifier for the collection task, and bind it with the data collection service ID to perform offline data collection tasks and obtain offline historical features;
[0011] The real-time data collection module is used to continuously trigger real-time data collection tasks in a real-time monitoring manner, generate a unique identifier for the collection task, and bind it with the data collection service ID to perform real-time data collection tasks and obtain real-time features;
[0012] The feature alignment module is used to receive offline historical features and real-time features respectively, perform feature alignment, form a cross-time scale feature matrix, and store it.
[0013] The offline data acquisition module includes a task scheduling unit, an offline data acquisition unit, and an offline data output unit; wherein:
[0014] The task scheduling unit is used to set the offline data collection time and trigger the offline data collection task;
[0015] The offline data acquisition unit is used to perform offline data acquisition tasks;
[0016] The offline data output unit is used to output offline historical features.
[0017] The offline data collection module further includes a service ID association unit; wherein:
[0018] The service ID association unit is used to generate a service ID for each offline data collection task.
[0019] The real-time data acquisition module includes a real-time monitoring unit, a real-time data acquisition unit, a feature calculation unit, and a real-time data output unit; wherein:
[0020] The real-time monitoring unit is used to trigger a real-time data collection task;
[0021] The real-time data acquisition unit is used to perform real-time data acquisition tasks;
[0022] The feature calculation unit is used to perform anomaly detection and time series prediction on the data obtained for the current task to obtain a real-time feature vector;
[0023] The real-time data output unit is used to output real-time features.
[0024] The feature alignment module includes a feature merging unit and a feature matrix forming unit; wherein:
[0025] The feature merging unit is used to receive offline historical features and real-time features and perform feature alignment;
[0026] The feature matrix forming unit is used to obtain and store a cross-time-scale feature matrix.
[0027] The risk warning module includes an offline risk scoring module, a real-time risk scoring module, and a risk fusion warning module; wherein:
[0028] The offline risk scoring module is used to establish an offline risk model and input an offline historical feature vector to obtain a first risk score;
[0029] The real-time risk scoring module is used to establish a real-time risk model and input a real-time feature vector to obtain a second risk score;
[0030] The risk fusion warning module is used to fuse the first risk score and the second risk score and output risk warning data.
[0031] The offline risk scoring module includes a first model building unit, a first feature vector input unit, and a first risk scoring unit; wherein:
[0032] The first model building unit is used to build an offline risk scoring model and train it using historical data;
[0033] The first feature vector input unit is used to load the offline risk scoring model and input the offline historical feature vector;
[0034] The first risk scoring unit is configured to obtain a first risk score for an offline historical feature vector.
[0035] The real-time risk scoring module includes a second model building unit, a second feature vector input unit, and a second risk scoring unit; wherein:
[0036] The second model building unit is used to build a real-time risk scoring model and train it using historical data;
[0037] The second feature vector input unit is used to load the real-time risk scoring model and input the real-time feature vector;
[0038] The second risk scoring unit is configured to obtain a second risk score for the real-time feature vector.
[0039] The monitoring point dynamic update module includes a key field extraction module and a monitoring point update module; wherein:
[0040] The key field extraction module is used to obtain risk warning data and extract key fields;
[0041] The monitoring point updating module is used to adjust the monitoring points of offline data and real-time data respectively according to key fields.
[0042] The present invention provides a system for constructing a digital risk prevention and control system for enterprises in the big health industry. The risk feature matrix establishment module is used to trigger offline data collection and real-time data collection requests, obtain offline historical features and real-time features, and construct a risk feature matrix across time scales; the risk warning module is used to establish offline risk models and real-time risk models respectively, and output risk warning data for data collection business IDs; the monitoring point dynamic update module is used to obtain risk warning data and dynamically update data monitoring points; by collecting offline data and real-time data respectively, the data monitoring points are dynamically updated according to the dual-channel correlation monitoring situation, thereby improving the accuracy of risk prevention and control. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0044] Figure 1 It is a structural principle diagram of the enterprise digital risk prevention and control system construction system of the big health industry of the present invention.
[0045] Figure 2 It is a flowchart of the steps of the method for constructing a digital risk prevention and control system for enterprises in the big health industry of the present invention.
[0046] Figure 3 It is a step flow chart of S100 of the present invention.
[0047] Figure 4 It is a step flow chart of S200 of the present invention.
[0048] Figure 5 It is a step flow chart of S300 of the present invention.
[0049] Figure 6 It is a structural principle diagram of the electronic device of the present invention.
[0050] 400-Risk feature matrix establishment module, 401-Task scheduling unit, 402-Offline data collection unit, 403-Business ID association unit, 404-Offline data output unit, 405-Real-time monitoring unit, 406-Real-time data collection unit, 407-Feature calculation unit, 408-Real-time data output unit, 409-Feature merging unit, 410-Feature matrix formation unit, 500-Risk warning module, 501-First model establishment unit, 502-First feature vector input unit, 503-First risk scoring unit, 504-Second model establishment unit, 505-Second feature vector input unit, 506-Second risk scoring unit, 507-Risk fusion warning module, 600-Monitoring point dynamic update module, 601-Key field extraction module, 602-Monitoring point update module. DETAILED DESCRIPTION
[0051] Exemplary embodiments are described in detail herein, with examples illustrated in the accompanying drawings. When the following description refers to the drawings, identical numerals in different drawings represent identical or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with this application.
[0052] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. As used in this application and the appended claims, the singular forms "a," "an," "the," and "the" are intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.
[0053] It should be understood that although the terms first, second, third, etc. may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".
[0054] See also Figure 1 The present invention provides a system for constructing a digital risk prevention and control system for an enterprise in the health industry, including a risk feature matrix establishment module 400, a risk warning module 500, and a monitoring point dynamic update module 600; wherein:
[0055] The risk feature matrix building module 400 is used to trigger offline data collection and real-time data collection requests, obtain offline historical features and real-time features, and build a risk feature matrix across time scales;
[0056] The risk warning module 500 is used to establish an offline risk model and a real-time risk model, and output risk warning data for the data collection business ID;
[0057] The monitoring point dynamic update module 600 is used to obtain risk warning data and dynamically update data monitoring points.
[0058] In this embodiment, the risk feature matrix establishment module 400 triggers offline data collection and real-time data collection requests, obtains offline historical features and real-time features, and constructs a risk feature matrix across time scales; the risk warning module 500 establishes offline risk models and real-time risk models respectively, and outputs risk warning data for the data collection business ID; the monitoring point dynamic update module 600 obtains risk warning data and dynamically updates data monitoring points; by collecting offline data and real-time data respectively, the data monitoring points are dynamically updated according to the dual-channel correlation monitoring situation, thereby improving the accuracy of risk prevention and control.
[0059] Furthermore, the risk feature matrix establishment module 400 includes an offline data acquisition module, a real-time data acquisition module, and a feature alignment module; wherein:
[0060] The offline data collection module is used to use task scheduling to regularly trigger offline data collection tasks, generate a unique identifier for the collection task, and bind it with the data collection service ID to perform offline data collection tasks and obtain offline historical features;
[0061] The real-time data collection module is used to continuously trigger real-time data collection tasks in a real-time monitoring manner, generate a unique identifier for the collection task, and bind it with the data collection service ID to perform real-time data collection tasks and obtain real-time features;
[0062] The feature alignment module is used to receive offline historical features and real-time features respectively, perform feature alignment, form a cross-time scale feature matrix, and store it.
[0063] In this embodiment, the offline data acquisition module adopts a task scheduling method to periodically trigger the offline data acquisition task, and generates a unique identifier for the acquisition task, and binds it with the data acquisition business ID to perform the offline data acquisition task and obtain offline historical features; the real-time data acquisition module adopts a real-time monitoring method to continuously trigger the real-time data acquisition task, and generates a unique identifier for the acquisition task, and binds it with the data acquisition business ID to perform the real-time data acquisition task and obtain real-time features; the feature alignment module receives offline historical features and real-time features respectively, performs feature alignment, forms a cross-time scale feature matrix, and stores it.
[0064] Furthermore, the offline data collection module includes a task scheduling unit 401, an offline data collection unit 402, a service ID association unit 403, and an offline data output unit 404; wherein:
[0065] The task scheduling unit 401 is used to set the offline data collection time and trigger the offline data collection task;
[0066] The offline data collection unit 402 is used to perform offline data collection tasks;
[0067] The business ID association unit 403 is used to generate a business ID for each offline data collection task;
[0068] The offline data output unit 404 is used to output offline historical features.
[0069] Furthermore, the real-time data acquisition module includes a real-time monitoring unit 405, a real-time data acquisition unit 406, a feature calculation unit 407, and a real-time data output unit 408; wherein:
[0070] The real-time monitoring unit 405 is used to trigger a real-time data collection task;
[0071] The real-time data acquisition unit 406 is used to perform real-time data acquisition tasks;
[0072] The feature calculation unit 407 is used to perform anomaly detection and time series prediction on the data obtained for the current task to obtain a real-time feature vector;
[0073] The real-time data output unit 408 is used to output real-time features.
[0074] Furthermore, the feature alignment module includes a feature merging unit 409 and a feature matrix forming unit 410; wherein:
[0075] The feature merging unit 409 is used to receive offline historical features and real-time features and perform feature alignment;
[0076] The feature matrix forming unit 410 is used to obtain and store a cross-time-scale feature matrix.
[0077] In this embodiment, the task scheduling unit 401 sets the offline data collection time and triggers the offline data collection task. The offline data collection task can be triggered regularly through the scheduling system (such as Airflow, DolphinScheduler), and the real-time data collection is triggered by monitoring business events (such as transaction occurrence, device status change) in real time through the message queue (such as Kafka). The offline data collection unit 402 executes the offline data collection task, and the acquisition method includes: reading historical data (such as user transaction records, equipment maintenance logs) from the enterprise data warehouse. Supplementary data (such as historical credit scores) is obtained from a third-party data platform (such as a credit reporting agency) through an API. The business ID association unit 403 generates a business ID for each offline data collection task, and at the same time generates a unique identifier and binds it to the business ID to ensure data traceability. The offline data output unit 404 outputs offline historical features; the real-time monitoring unit 405 triggers the real-time data collection task; the real-time data collection unit 406 executes the real-time data collection task; the feature calculation unit 407 is used to perform anomaly detection and time series prediction on the data obtained for the current task, and obtain real-time feature vectors, wherein statistical features include: calculating the average of the user's historical transaction amount, the equipment failure rate, etc. Time series features: extracting the user transaction frequency trend and the equipment operation time distribution. Text features: performing word segmentation and TF-IDF vectorization on the log text (such as NLP processing of customer service feedback). The real-time data output unit 408 outputs real-time features; the feature merging unit 409 receives offline historical features and real-time features, and performs feature alignment; the feature matrix forming unit 410 obtains and stores the cross-time scale feature matrix.
[0078] Furthermore, the risk warning module 500 includes an offline risk scoring module, a real-time risk scoring module, and a risk fusion warning module 507; wherein:
[0079] The offline risk scoring module is used to establish an offline risk model and input an offline historical feature vector to obtain a first risk score;
[0080] The real-time risk scoring module is used to establish a real-time risk model and input a real-time feature vector to obtain a second risk score;
[0081] The risk fusion warning module 507 is used to fuse the first risk score and the second risk score to output risk warning data.
[0082] In this embodiment, the offline risk scoring module establishes an offline risk model and inputs an offline historical feature vector to obtain a first risk score.
[0083] For example, the logistic regression model. The formula of the logistic regression model is:
[0084]
[0085] Among them, y is the risk label (y = 1 means risk, y = 0 means no risk), x is the offline historical feature vector, w is the weight vector, and b is the bias term.
[0086] The logistic regression model is trained using training data, where the training data (including historical features and corresponding risk labels) is used to obtain the optimal w and b.
[0087] The offline historical feature vector x offline Input into the offline risk model. Calculate the first risk score offline_risk_score based on the input feature vector, that is, P(y=1|x offline ).
[0088] The real-time risk scoring module establishes a real-time risk model and inputs a real-time feature vector to obtain a second risk score.
[0089] Use real-time data and real-time features to build a real-time risk model, such as a sliding window model combined with simple rule judgment. Use a time series-based model (such as an ARIMA model) to predict real-time risk trends. Input the real-time feature vector xreal_time into the real-time risk model. The model calculates a second risk score real_time_risk_score based on the input feature vector. For example, based on the rule judgment result, if the rule condition is met, real_time_risk_score = 0.8, otherwise real_time_risk_score = 0.2
[0090] The risk fusion warning module 507 fuses the first risk score and the second risk score to output risk warning data. The first risk score offline_risk_score and the second risk score real_time_risk_score are fused, for example, using a weighted average method:
[0091] final_risk_score=α×offline_risk_score+(1-α)×real_time_risk_score;
[0092] Among them, α is the weight of offline risk score, 0≤α≤1, and can be adjusted according to actual business needs.
[0093] Generate risk warning data based on the comprehensive risk score final_risk_score. For example, set a threshold threshold. If final_risk_score > threshold, then output risk warning data, including information such as the data collection business ID, risk score, and risk type.
[0094] Furthermore, the offline risk scoring module includes a first model building unit 501, a first feature vector input unit 502, and a first risk scoring unit 503; wherein:
[0095] The first model building unit 501 is used to build an offline risk scoring model and train it using historical data;
[0096] The first feature vector input unit 502 is used to load the offline risk scoring model and input the offline historical feature vector;
[0097] The first risk scoring unit 503 is configured to obtain a first risk score for the offline historical feature vector.
[0098] Furthermore, the real-time risk scoring module includes a second model building unit 504, a second feature vector input unit 505, and a second risk scoring unit 506; wherein:
[0099] The second model building unit 504 is used to build a real-time risk scoring model and train it using historical data;
[0100] The second feature vector input unit 505 is used to load the real-time risk scoring model and input the real-time feature vector;
[0101] The second risk scoring unit 506 is configured to obtain a second risk score for the real-time feature vector.
[0102] In this embodiment, the first model building unit 501 builds an offline risk scoring model, wherein the model includes: time series anomaly detection, using Isolation Forest to detect real-time transaction anomalies; time series prediction, using Prophet to predict the temperature change trend of the equipment, and using historical data to train it; the first feature vector input unit 502 loads the offline risk scoring model and inputs the offline historical feature vector; the first risk scoring unit 503 obtains a first risk score based on the offline historical feature vector; the second model building unit 504 is used to establish a real-time risk scoring model and train it using historical data; the second feature vector input unit 505 loads the real-time risk scoring model and inputs the real-time feature vector; the second risk scoring unit 506 obtains a second risk score based on the real-time feature vector.
[0103] Furthermore, the monitoring point dynamic update module 600 includes a key field extraction module 601 and a monitoring point update module 602; wherein:
[0104] The key field extraction module 601 is used to obtain risk warning data and extract key fields;
[0105] The monitoring point updating module 602 is used to adjust the monitoring points of offline data and real-time data respectively according to key fields.
[0106] In this embodiment, the key field extraction module 601 obtains risk warning data and extracts key fields; the monitoring point update module 602 adjusts the monitoring points of offline data and real-time data respectively according to the key fields.
[0107] Corresponding to the aforementioned embodiment of the system for constructing a digital risk prevention and control system for an enterprise in the big health industry, the present application also provides an embodiment of a method for constructing a digital risk prevention and control system for an enterprise in the big health industry.
[0108] Please refer to Figures 2 to 5 , the method may include the following steps:
[0109] S100: Trigger offline data collection and real-time data collection requests, obtain offline historical features and real-time features, and build a risk feature matrix across time scales;
[0110] S200: Establish an offline risk model and a real-time risk model respectively, and output risk warning data based on the data collection business ID;
[0111] S300: Obtain risk warning data and dynamically update data monitoring points.
[0112] In this embodiment, offline data collection and real-time data collection requests are first triggered to obtain offline historical features and real-time features, and a risk feature matrix across time scales is constructed; then, offline risk models and real-time risk models are established respectively, and risk warning data is output for the data collection business ID; finally, risk warning data is obtained and data monitoring points are dynamically updated; by collecting offline data and real-time data respectively, data monitoring points are dynamically updated according to the dual-channel correlation monitoring situation, thereby improving the accuracy of risk prevention and control.
[0113] Among them, in the steps of triggering offline data collection and real-time data collection requests, obtaining offline historical features and real-time features, and constructing a risk feature matrix across time scales, the specific process is as follows:
[0114] S101: Using task scheduling, regularly trigger offline data collection tasks, generate a unique identifier for the collection task, and bind it to the data collection service ID to perform offline data collection tasks and obtain offline historical features;
[0115] S102: Using real-time monitoring mode, continuously triggering real-time data collection tasks, generating a unique identifier for the collection task, and binding it with the data collection service ID to perform real-time data collection tasks and obtain real-time features;
[0116] S103: Receive offline historical features and real-time features respectively, perform feature alignment, form a cross-time scale feature matrix, and store it.
[0117] Among them, in the steps of establishing the offline risk model and the real-time risk model respectively, outputting the risk warning data for the data collection business ID, the specific process is as follows:
[0118] S201: Establish an offline risk model and input an offline historical feature vector to obtain a first risk score;
[0119] S202: Establish a real-time risk model and input a real-time feature vector to obtain a second risk score;
[0120] S203: Integrate the first risk score and the second risk score to output risk warning data.
[0121] Among them, in the steps of obtaining risk warning data and dynamically updating data monitoring points, the specific process is as follows:
[0122] S301: Obtain risk warning data and extract key fields;
[0123] S302: Adjust the monitoring points of offline data and real-time data respectively according to the key fields.
[0124] Regarding the method in the above embodiment, the specific modules for executing operations in each process have been described in detail in the embodiment of the system and will not be elaborated here.
[0125] As for the method embodiment, since it basically corresponds to the system embodiment, the relevant parts can be referred to the partial description of the system embodiment.
[0126] Accordingly, the present application also provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned method for constructing a digital risk prevention and control system for enterprises in the big health industry. Figure 6As shown in the figure, a hardware structure diagram of any device with data processing capability in a system for building a digital risk prevention and control system for a large health industry provided by an embodiment of the present invention is provided. Figure 6 In addition to the processor, memory, and network interface shown, any device with data processing capabilities in which the apparatus in the embodiment is located may also include other hardware, generally based on the actual functions of the device with data processing capabilities, which will not be described in detail.
[0127] Accordingly, the present application also provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the method for constructing a digital risk prevention and control system for an enterprise in the big health industry as described above. The computer-readable storage medium may be an internal storage unit of any device with data processing capabilities as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium may also be an external storage device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), an SD card, a flash card (Flash Card), etc. equipped on the device. Furthermore, the computer-readable storage medium may also include both an internal storage unit and an external storage device of any device with data processing capabilities. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capabilities, and may also be used to temporarily store data that has been output or is to be output.
[0128] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the contents disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of this application and include common knowledge or customary techniques in the art that are not disclosed in this application.
[0129] It will be understood that the present application is not limited to the exact construction that has been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof.
Claims
1. A system for building a digital risk prevention and control system for enterprises in the health industry, characterized by: It includes risk feature matrix establishment module, risk warning module, and monitoring point dynamic update module; among which: The risk feature matrix establishment module is used to trigger offline data collection and real-time data collection requests, obtain offline historical features and real-time features, and construct a risk feature matrix across time scales; The risk warning module is used to establish an offline risk model and a real-time risk model, and output risk warning data for the data collection business ID; The monitoring point dynamic update module is used to obtain risk warning data and dynamically update data monitoring points.
2. The enterprise digital risk prevention and control system construction system for the big health industry according to claim 1 is characterized in that: The risk feature matrix establishment module includes an offline data acquisition module, a real-time data acquisition module, and a feature alignment module; wherein: The offline data collection module is used to use task scheduling to regularly trigger offline data collection tasks, generate a unique identifier for the collection task, and bind it with the data collection service ID to perform offline data collection tasks and obtain offline historical features; The real-time data collection module is used to continuously trigger real-time data collection tasks in a real-time monitoring manner, generate a unique identifier for the collection task, and bind it with the data collection service ID to perform real-time data collection tasks and obtain real-time features; The feature alignment module is used to receive offline historical features and real-time features respectively, perform feature alignment, form a cross-time scale feature matrix, and store it.
3. The enterprise digital risk prevention and control system construction system for the big health industry as claimed in claim 2 is characterized in that: The offline data acquisition module includes a task scheduling unit, an offline data acquisition unit, and an offline data output unit; wherein: The task scheduling unit is used to set the offline data collection time and trigger the offline data collection task; The offline data acquisition unit is used to perform offline data acquisition tasks; The offline data output unit is used to output offline historical features.
4. The enterprise digital risk prevention and control system construction system for the big health industry according to claim 3 is characterized in that: The offline data collection module also includes a business ID association unit; wherein: The service ID association unit is used to generate a service ID for each offline data collection task.
5. The enterprise digital risk prevention and control system construction system for the big health industry according to claim 4 is characterized in that: The real-time data acquisition module includes a real-time monitoring unit, a real-time data acquisition unit, a feature calculation unit, and a real-time data output unit; wherein: The real-time monitoring unit is used to trigger a real-time data collection task; The real-time data acquisition unit is used to perform real-time data acquisition tasks; The feature calculation unit is used to perform anomaly detection and time series prediction on the data obtained for the current task to obtain a real-time feature vector; The real-time data output unit is used to output real-time features.
6. The enterprise digital risk prevention and control system construction system for the big health industry according to claim 5 is characterized in that: The feature alignment module includes a feature merging unit and a feature matrix forming unit; wherein: The feature merging unit is used to receive offline historical features and real-time features and perform feature alignment; The feature matrix forming unit is used to obtain and store a cross-time-scale feature matrix.
7. The enterprise digital risk prevention and control system construction system for the big health industry according to claim 1 is characterized in that: The risk warning module includes an offline risk scoring module, a real-time risk scoring module, and a risk fusion warning module; wherein: The offline risk scoring module is used to establish an offline risk model and input an offline historical feature vector to obtain a first risk score; The real-time risk scoring module is used to establish a real-time risk model and input a real-time feature vector to obtain a second risk score; The risk fusion warning module is used to fuse the first risk score and the second risk score and output risk warning data.
8. The enterprise digital risk prevention and control system construction system for the big health industry according to claim 7 is characterized in that: The offline risk scoring module includes a first model building unit, a first feature vector input unit, and a first risk scoring unit; wherein: The first model building unit is used to build an offline risk scoring model and train it using historical data; The first feature vector input unit is used to load the offline risk scoring model and input the offline historical feature vector; The first risk scoring unit is configured to obtain a first risk score for an offline historical feature vector.
9. The enterprise digital risk prevention and control system construction system for the big health industry according to claim 7 is characterized in that: The real-time risk scoring module includes a second model building unit, a second feature vector input unit, and a second risk scoring unit; wherein: The second model building unit is used to build a real-time risk scoring model and train it using historical data; The second feature vector input unit is used to load the real-time risk scoring model and input the real-time feature vector; The second risk scoring unit is configured to obtain a second risk score for the real-time feature vector.
10. The enterprise digital risk prevention and control system construction system for the big health industry according to claim 1 is characterized in that: The monitoring point dynamic update module includes a key field extraction module and a monitoring point update module; wherein: The key field extraction module is used to obtain risk warning data and extract key fields; The monitoring point updating module is used to adjust the monitoring points of offline data and real-time data respectively according to key fields.