Real-time credit risk monitoring decision platform based on AI

Abstract

The invention discloses a real-time credit and loan risk monitoring decision platform based on AI, relates to the technical field of risk monitoring, and discloses a credit and loan data monitoring module, a credit and loan client monitoring module, a credit and loan risk analysis module, a risk client rating module and a monitoring decision updating module. The credit data monitoring module, the credit customer monitoring module, the credit risk analysis module and the risk customer rating module are arranged, so that the credit monitoring data of the credit customer can be monitored in real time, the risk analysis of the credit customer is completed by a reasonable monitoring strategy, whether the credit customer is a risk customer or not is efficiently judged, and the risk of the credit customer is effectively judged. The credit risk of the risk customer is comprehensively and deeply analyzed, the credit personnel can conveniently execute corresponding management and operation through the level division of the credit risk, and the credit risk monitoring decision of the platform can be updated by setting the risk customer rating module, so that the financial institution can quickly respond to the change of the market and the change of the customer risk condition.

Description

Technical Field

[0001] The present invention relates to the technical field of risk monitoring, and more specifically, to an AI-based real-time credit risk monitoring decision-making platform. Background Art

[0002] In the current financial industry, credit business is one of the core businesses, and its risk management and monitoring are crucial to the sound operation of financial institutions. With the rapid development of technologies such as big data and artificial intelligence, financial institutions have begun to explore the use of these new technologies to optimize the credit risk monitoring process and improve the accuracy and timeliness of risk management.

[0003] Although some AI-based credit risk monitoring systems have been developed, these systems still have some shortcomings. For example, the system lacks depth in risk analysis and determination, and cannot accurately identify potential risk points. The system needs to analyze all credit data of credit customers before determining whether there is credit risk, and the efficiency of credit risk determination is not high. In addition, the market environment and customer risk status are constantly changing, and traditional credit risk monitoring methods often find it difficult to quickly adapt to these changes. Financial institutions need a system that can update monitoring strategies in real time and dynamically optimize monitoring decisions to cope with market changes and changes in customer risk status. Summary of the invention

[0004] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a real-time credit risk monitoring decision-making platform based on AI.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] AI-based real-time credit risk monitoring and decision-making platform, including credit data monitoring module, credit customer monitoring module, credit risk analysis module, risk customer rating module, and monitoring decision update module;

[0007] The credit data monitoring module is used to determine the credit customers of the financial institution, monitor the credit customers, and collect various types of credit monitoring data of the credit customers in real time;

[0008] The credit customer monitoring module is used to set a monitoring decision cycle, and each time a cycle node of the monitoring decision cycle is passed, risk monitoring is performed on each credit customer in sequence according to the credit decision order;

[0009] The credit risk analysis module obtains the risk data analysis order of the credit customer each time risk monitoring is performed on a credit customer, analyzes various types of credit monitoring data of the credit customer in sequence according to the risk data analysis order, generates a credit data analysis log each time a type of credit monitoring data of a credit customer is analyzed, obtains credit data analysis parameters of the credit data analysis log, and determines whether to mark the credit customer as a risky customer based on a comparison result of the credit data analysis parameters and the credit data analysis threshold parameters;

[0010] Whenever the risk customer rating module completes the analysis of all types of credit monitoring data of risk customers, it obtains the credit risk index of the risk customer, determines the credit risk level of the risk customer based on the credit risk index, and simultaneously generates the credit risk record of the risk customer;

[0011] The monitoring decision updating module updates the credit decision order and the risk data parsing sequence of each credit customer when risk monitoring of all credit customers in the credit decision order is completed. The updated credit decision order is used for risk monitoring at the next cycle node.

[0012] Furthermore, the credit data analysis log includes the credit customer ID, the credit monitoring data type, the credit data analysis parameters, and the data analysis time.

[0013] Furthermore, the credit data parsing parameters of the credit data parsing log are obtained through the following steps: collecting a type of credit monitoring data of the credit customer in the previous monitoring decision cycle, obtaining the credit data parsing model corresponding to the type of credit monitoring data, preprocessing the credit monitoring data, and then inputting the credit monitoring data into the credit data parsing model, and the credit data parsing model outputs the credit data parsing parameters.

[0014] Further, based on the comparison result between the credit data parsing parameter and the credit data parsing threshold parameter, it is determined whether to mark the credit customer as a risk customer, specifically: the credit data parsing threshold parameter is set, when the credit data parsing parameter of the credit data parsing log is greater than or equal to the credit data parsing threshold parameter, the type of credit monitoring data of the credit data parsing log is obtained, the type of credit monitoring data is marked as credit risk data, and the credit customer is marked as a risk customer, and then the next type of credit monitoring data of the risk customer is parsed according to the risk data parsing order;

[0015] When the credit data parsing parameter of the credit data parsing log is less than the credit data parsing threshold parameter, the next type of credit monitoring data of the credit customer is parsed in the risk data parsing order.

[0016] Furthermore, the credit risk record includes the credit risk level, the credit risk index, and the risk record time.

[0017] Furthermore, the credit risk index of the risky customer is obtained by the following steps: obtaining all the credit risk data of the risky customer, matching all the credit risk data into a credit risk group, when two credit risk data in the credit risk group are adjacent in the risk data parsing order, marking the credit risk group as a risk continuous group, when two credit risk data in the credit risk group are not adjacent in the risk data parsing order, no further processing is performed, marking the total number of risk continuous groups as RiskDP, marking the total number of credit risk data as RiskAL, and using the formula The credit risk index i nde of the risky customer is obtained, where u1 is the risk continuity coefficient.

[0018] Furthermore, the credit decision order is updated and obtained through the following steps: obtaining the credit decision value of each credit customer, sorting all credit customers in descending order according to the numerical value of the credit decision value, and generating the credit decision order according to the sorting result.

[0019] Furthermore, the credit decision value of the credit customer is obtained by the following steps: obtain all credit risk records of the credit customer within N time periods, mark the total number of credit risk records as RecoNs, sort them in the order of risk record time, sum the credit risk indexes of two adjacent credit risk records after sorting to obtain the credit risk superposition index, sum all credit risk superposition indexes and take the average to obtain the average risk superposition index Javp, and use the formula Get the credit decision value XDJ of the credit customer.

[0020] Furthermore, the risk data parsing order of the credit customer is updated and obtained through the following steps: obtaining all credit data parsing logs of the risk customer within time c, obtaining the credit monitoring data type of the credit data parsing log, marking the credit data parsing logs of the same type of credit monitoring data as the same type of data parsing logs, and then obtaining the credit data parsing values ​​of each type of credit monitoring data, sorting each type of credit monitoring data in descending order according to the numerical value of the credit data parsing value, and generating the risk data parsing order based on the sorting results.

[0021] Further, the credit data parsing value of a type of credit monitoring data is obtained by the following steps: summing up and averaging the credit data parsing parameters of all similar data parsing logs to obtain an average parsing parameter Jexcm; sorting all similar data parsing logs in the order of data parsing time; taking the difference between the credit data parsing parameters of two adjacent similar data parsing logs after sorting and taking the absolute value to obtain a parsing parameter floating value; summing up and averaging all parsing parameter floating values ​​to obtain an average parsing parameter floating value F l oaB; using the formula The credit data analysis value Kew of this type of credit monitoring data is obtained, where p1 is the analysis parameter coefficient and p2 is the parameter floating coefficient.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] The credit data monitoring module, credit customer monitoring module, credit risk analysis module and risk customer rating module are set up to monitor the credit monitoring data of credit customers in real time, and complete the risk analysis of credit customers with reasonable monitoring strategies, effectively determine whether credit customers are risky customers, comprehensively and deeply analyze the credit risks of risky customers, and facilitate credit personnel to perform corresponding management and operations through the classification of credit risks;

[0024] Setting up a risk customer rating module can update the platform's credit risk monitoring decisions, allowing financial institutions to quickly respond to market changes and changes in customer risk conditions, and dynamically optimize monitoring decisions on credit customers and credit monitoring data. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is the operation flow chart of the AI-based real-time credit risk monitoring decision platform;

[0026] Figure 2 A flow chart of the method for determining risky customers;

[0027] Figure 3 This is a module block diagram of the AI-based real-time credit risk monitoring decision-making platform. DETAILED DESCRIPTION

[0028] Reference Figures 1 to 3 , an AI-based real-time credit risk monitoring and decision-making platform, including credit data monitoring module, credit customer monitoring module, credit risk analysis module, risk customer rating module, and monitoring decision update module.

[0029] Credit data monitoring module: Identify the credit customers of financial institutions, monitor the credit customers, and collect various types of credit monitoring data of credit customers in real time (credit monitoring data refers to the relevant data that financial institutions need to monitor and use to analyze whether credit customers are credit risks during the lending process. The data types of credit monitoring data include but are not limited to financial data, operating data, market behavior data, and credit data).

[0030] Credit customer monitoring module: set the monitoring decision cycle, and conduct risk monitoring on each credit customer in the credit decision order at each cycle node of the monitoring decision cycle (the credit decision order includes all credit customers of the financial institution, and arranges the credit customers in order. For example, if the institution includes a total of 4 credit customers, namely credit customer No. 1, credit customer No. 2, credit customer No. 3, and credit customer No. 4, the credit decision order is 1, credit customer No. 2; 2, credit customer No. 3; 3, credit customer No. 1; 4, credit customer No. 4, then the credit customers are monitored for risk in order).

[0031] Credit risk analysis module: Each time a credit customer is monitored for risk, the risk data analysis sequence of the credit customer is obtained (the risk data analysis sequence includes all types of credit monitoring data, and each type of credit monitoring data is arranged in sequence. For example, there are three types of credit monitoring data in total, namely financial data, operating data, and market behavior data. The risk data analysis sequence of the credit customer is 1. operating data; 2. market behavior data; 3. financial data, then each type of credit monitoring data of the credit customer is analyzed in sequence). Each type of credit monitoring data of the credit customer is analyzed in sequence according to the risk data analysis sequence. When each type of credit monitoring data of a credit customer is analyzed, a credit risk analysis module is generated. Data analysis log, the credit data analysis log includes the credit customer ID, the type of credit monitoring data, the credit data analysis parameters, and the data analysis time. The credit data analysis parameters of the credit data analysis log are obtained, and the credit data analysis threshold parameters are set (the credit data analysis threshold parameters are the set values ​​of the platform). When the credit data analysis parameters of the credit data analysis log are greater than or equal to the credit data analysis threshold parameters, the credit monitoring data type of the credit data analysis log is obtained, and the credit monitoring data of this type is marked as credit risk data, and the credit customer is marked as a risk customer (if the customer has been marked as a risk customer, there is no need to mark it again), and then the next type of credit monitoring data of the risk customer is analyzed in the risk data analysis order.

[0032] When the credit data parsing parameter of the credit data parsing log is less than the credit data parsing threshold parameter, the next type of credit monitoring data of the credit customer is parsed in the risk data parsing order.

[0033] The credit data parsing parameters of the credit data parsing log are obtained through the following steps: collecting a type of credit monitoring data of the credit customer in the previous monitoring decision cycle, obtaining the credit data parsing model corresponding to the type of credit monitoring data, preprocessing the credit monitoring data (preprocessing methods include data cleaning, data normalization, outlier processing, missing value processing, etc.), and then inputting the credit monitoring data into the credit data parsing model, and the credit data parsing model outputs the credit data parsing parameters.

[0034] Each type of credit monitoring data corresponds to a credit data analysis model. Each credit data analysis model is constructed based on a neural network model. The only difference between each neural network model is the replacement of training data. This embodiment introduces the construction steps of a credit data analysis model for financial data by taking L groups of financial data in the monitoring decision cycle as an example. (If it is a credit data analysis model for operating data, then L groups of operating data in the monitoring decision cycle are collected.) Each group of financial data includes operating income, net profit, gross profit margin, debt-to-asset ratio and other data. The financial data can be historical data collected in real life or data set for simulation of neural network model training. A neural network model is constructed and the financial data is used as training data for the neural network model. The training data is assigned credit data parsing parameters, and the value range of the credit data parsing parameters is (0.1-3.9). The closer the value of the credit data parsing parameter is to 3.9, the greater the credit risk of the financial data, and the closer the value of the credit data parsing parameter is to 0.1, the smaller the credit risk of the financial data (if it is a credit data parsing model for operating data, the closer the value of the credit data parsing parameter is to 3.9, the greater the credit risk of the operating data, and the closer the value of the credit data parsing parameter is to 0.1, the smaller the credit risk of the operating data). The training data is divided into 70% training set, 15% validation set and 15% test set, and the neural network model is iteratively trained to obtain a credit data parsing model for financial data.

[0035] By setting up a credit data monitoring module, a credit customer monitoring module, a credit risk analysis module and a risky customer rating module, it is possible to monitor the credit monitoring data of credit customers in real time, complete the risk analysis of credit customers with reasonable monitoring strategies, efficiently determine whether a credit customer is a risky customer, comprehensively and deeply analyze the credit risk of risky customers, and facilitate credit personnel to perform corresponding management and operations through the classification of credit risks.

[0036] Risky customer rating module: When all types of credit monitoring data of risky customers are parsed (i.e., all types of credit monitoring data in the risk data parsing sequence are parsed in sequence), the credit risk index of the risky customer is obtained, the credit risk level of the risky customer is determined based on the credit risk index, and the credit risk record of the risky customer is generated simultaneously. The credit risk record includes the credit risk level, credit risk index, and risk record time. Credit personnel can take corresponding measures against risky customers according to the credit risk level.

[0037] The credit risk index of risky customers is obtained through the following steps: obtain all credit risk data of risky customers (limited to credit risk data analyzed after the cycle node), match all credit risk data into a credit risk group, and when two credit risk data in a credit risk group are adjacent in the risk data parsing order (for example, two credit risk data in a credit risk group are market behavior data and financial data, and the risk data parsing order of the credit customer is 1, operating data; 2, market behavior data; 3, financial data, indicating that the two credit risk data are adjacent in the risk data parsing order), mark the credit risk group as a risk continuity group. When two credit risk data in a credit risk group are not adjacent in the risk data parsing order, no further processing is performed, and the total number of risk continuity groups is marked as RiskDP, and the total number of credit risk data is marked as RiskAL. The credit risk index i nde of risky customers is obtained, where u1 is the risk continuity coefficient and the value of u1 is 0.84.

[0038] Determining the credit risk level of a risky customer based on the credit risk index is specifically as follows: obtaining the credit risk index i nde of the risky customer, wherein the range of each credit risk index i nde corresponds to a credit risk level, and the range of the credit risk index i nde is [0, i nd1], (i nd1, i nd2], …, (i nde-1, i nde], and the credit risk levels include credit risk level 1, credit risk level 2, …, credit risk level e-1, and credit risk level e, and the credit risk of credit risk level 1 is less than the credit risk of credit risk level 2, …, less than the credit risk of credit risk level e-1, and less than the credit risk of credit risk level e.

[0039] Monitoring decision update module: When all credit customers within the credit decision sequence complete risk monitoring, the credit decision sequence and the risk data analysis order of each credit customer are updated. The updated credit decision sequence is used for risk monitoring at the next cycle node.

[0040] The credit decision order is updated and obtained through the following steps: obtaining the credit decision value of each credit customer, sorting all credit customers in descending order according to the numerical value of the credit decision value, and generating the credit decision order based on the sorting result.

[0041] The credit decision value of a credit customer is obtained through the following steps: obtain all credit risk records of the credit customer within N time periods, mark the total number of credit risk records as RecoNs, sort them in the order of risk record time, sum the credit risk indexes of two adjacent credit risk records after sorting to obtain the credit risk superposition index, sum all credit risk superposition indexes and take the average to obtain the average risk superposition index Javp, and use the formula Get the credit decision value XDJ of the credit customer.

[0042] The risk data parsing order of a credit customer is updated and obtained through the following steps: obtaining all credit data parsing logs of the risk customer within time c, obtaining the credit monitoring data type of the credit data parsing log, marking the credit data parsing logs of the same type of credit monitoring data as the same type of data parsing logs, and then obtaining the credit data parsing values ​​of each type of credit monitoring data, sorting each type of credit monitoring data in descending order according to the numerical value of the credit data parsing value, and generating the risk data parsing order based on the sorting results.

[0043] The credit data parsing value of a type of credit monitoring data is obtained by the following steps: summing up and averaging the credit data parsing parameters of all similar data parsing logs to obtain the average parsing parameter Jexcm; sorting all similar data parsing logs in the order of data parsing time; taking the difference between the credit data parsing parameters of two adjacent similar data parsing logs after sorting and taking the absolute value to obtain the parsing parameter floating value; summing up and averaging all the parsing parameter floating values ​​to obtain the average parsing parameter floating value F l oaB; using the formula The credit data analysis value Kew of this type of credit monitoring data is obtained, where p1 is the analysis parameter coefficient, p2 is the parameter floating coefficient, the value of p1 is 1.28, and the value of p2 is 1.33.

[0044] Setting up a risk customer rating module can update the platform's credit risk monitoring decisions, allowing financial institutions to quickly respond to market changes and changes in customer risk conditions, and dynamically optimize monitoring decisions on credit customers and credit monitoring data.

[0045] The above formulas are all dimensionless and numerical calculations are performed, and the preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0046] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0047] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0048] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0049] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0050] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0051] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0052] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. AI-based real-time credit risk monitoring and decision-making platform, characterized by: It includes credit data monitoring module, credit customer monitoring module, credit risk analysis module, risk customer rating module, and monitoring decision update module; The credit data monitoring module is used to determine the credit customers of the financial institution, monitor the credit customers, and collect various types of credit monitoring data of the credit customers in real time; The credit customer monitoring module is used to set a monitoring decision cycle, and each time a cycle node of the monitoring decision cycle is passed, risk monitoring is performed on each credit customer in sequence according to the credit decision order; The credit risk analysis module obtains the risk data analysis order of the credit customer each time risk monitoring is performed on a credit customer, analyzes various types of credit monitoring data of the credit customer in sequence according to the risk data analysis order, generates a credit data analysis log each time a type of credit monitoring data of a credit customer is analyzed, obtains credit data analysis parameters of the credit data analysis log, and determines whether to mark the credit customer as a risky customer based on a comparison result of the credit data analysis parameters and the credit data analysis threshold parameters; Whenever the risk customer rating module completes the analysis of all types of credit monitoring data of risk customers, it obtains the credit risk index of the risk customer, determines the credit risk level of the risk customer based on the credit risk index, and simultaneously generates the credit risk record of the risk customer; The monitoring decision updating module updates the credit decision order and the risk data parsing sequence of each credit customer when risk monitoring of all credit customers in the credit decision order is completed. The updated credit decision order is used for risk monitoring at the next cycle node.

2. The AI-based real-time credit risk monitoring and decision-making platform according to claim 1 is characterized in that: The credit data analysis log includes the credit customer ID, credit monitoring data type, credit data analysis parameters, and data analysis time.

3. The AI-based real-time credit risk monitoring and decision-making platform according to claim 1 is characterized in that: The credit data parsing parameters of the credit data parsing log are obtained through the following steps: collecting a type of credit monitoring data of the credit customer in the previous monitoring decision cycle, obtaining the credit data parsing model corresponding to the type of credit monitoring data, preprocessing the credit monitoring data, and then inputting the credit monitoring data into the credit data parsing model, and the credit data parsing model outputs the credit data parsing parameters.

4. The AI-based real-time credit risk monitoring and decision-making platform according to claim 1 is characterized in that: Based on the comparison result of the credit data parsing parameter and the credit data parsing threshold parameter, determine whether to mark the credit customer as a risk customer, specifically: set the credit data parsing threshold parameter, when the credit data parsing parameter of the credit data parsing log is greater than or equal to the credit data parsing threshold parameter, obtain the credit monitoring data type of the credit data parsing log, mark the type of credit monitoring data as credit risk data, and mark the credit customer as a risk customer, and then parse the next type of credit monitoring data of the risk customer according to the risk data parsing order; When the credit data parsing parameter of the credit data parsing log is less than the credit data parsing threshold parameter, the next type of credit monitoring data of the credit customer is parsed in the risk data parsing order.

5. The AI-based real-time credit risk monitoring and decision-making platform according to claim 1 is characterized in that: Credit risk records include credit risk level, credit risk index and risk record time.

6. The AI-based real-time credit risk monitoring and decision-making platform according to claim 1, characterized in that: The credit risk index of risky customers is obtained by the following steps: obtain all credit risk data of risky customers, match all credit risk data into a credit risk group, and when two credit risk data in a credit risk group are adjacent in the risk data parsing order, mark the credit risk group as a risk continuous group. When two credit risk data in a credit risk group are not adjacent in the risk data parsing order, no further processing is performed, and the total number of risk continuous groups is marked as RiskDP, and the total number of credit risk data is marked as RiskAL. Using the formula Get the credit risk index inde of the risky customer, where u1 is the risk continuity coefficient.

7. The AI-based real-time credit risk monitoring and decision-making platform according to claim 1, characterized in that: The credit decision order is updated and obtained through the following steps: obtaining the credit decision value of each credit customer, sorting all credit customers in descending order according to the numerical value of the credit decision value, and generating the credit decision order based on the sorting result.

8. The AI-based real-time credit risk monitoring and decision-making platform according to claim 7 is characterized in that: The credit decision value of a credit customer is obtained through the following steps: obtain all credit risk records of the credit customer within N time periods, mark the total number of credit risk records as RecoNs, sort them in the order of risk record time, sum the credit risk indexes of two adjacent credit risk records after sorting to obtain the credit risk superposition index, sum all credit risk superposition indexes and take the average to obtain the average risk superposition index Javp, and use the formula Get the credit decision value XDJ of the credit customer.

9. The AI-based real-time credit risk monitoring and decision-making platform according to claim 1, characterized in that: The risk data parsing order of a credit customer is updated and obtained through the following steps: obtaining all credit data parsing logs of the risk customer within time c, obtaining the credit monitoring data type of the credit data parsing log, marking the credit data parsing logs of the same type of credit monitoring data as the same type of data parsing logs, and then obtaining the credit data parsing values ​​of each type of credit monitoring data, sorting each type of credit monitoring data in descending order according to the numerical value of the credit data parsing value, and generating the risk data parsing order based on the sorting results.

10. The AI-based real-time credit risk monitoring and decision-making platform according to claim 9, characterized in that: The credit data analysis value of a type of credit monitoring data is obtained by the following steps: summing up and averaging the credit data analysis parameters of all similar data analysis logs to obtain the average analysis parameter Jexcm; sorting all similar data analysis logs in the order of data analysis time; taking the difference between the credit data analysis parameters of two adjacent similar data analysis logs after sorting and taking the absolute value to obtain the analysis parameter floating value; summing up and averaging all analysis parameter floating values ​​to obtain the average analysis parameter floating value FloaB; using the formula The credit data analysis value Kew of this type of credit monitoring data is obtained, where p1 is the analysis parameter coefficient and p2 is the parameter floating coefficient.

Images