Automobile financial credit AI robot risk control method, device, equipment and medium
Through the AI robot risk control method, multi-level credit inquiry and strategy set analysis, the problems of low approval efficiency and risk deviation caused by manual review in automobile financial credit are solved, and accurate judgment and second-level response to customers' potential risks are achieved. Improved approval efficiency and risk prevention capabilities.
Patent Information
- Application Number
- CN202411098745.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing auto finance credit institutions rely on manual review and are difficult to form a comprehensive and systematic risk assessment, resulting in low approval efficiency and deviations from credit risk and approval amount.
Using AI robot risk control method, through multi-level credit query and intelligent analysis of strategy sets, we obtain customer star classification, and combine vehicle financing information to perform intelligent analysis of multi-level function sets, select strategy mode for risk decision diversion processing, and realize customer Accurate judgment and second-level response to potential credit risks.
It improves the approval efficiency of auto finance credit applications, achieves second-level response, accurately judges customer risk levels and types, optimizes customer experience, and enhances risk prevention capabilities through real-time monitoring and model optimization.
Smart Images

Figure CN119090621A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of automobile financial credit risk prevention and control analysis, and in particular to an automobile financial credit AI robot risk control method, device, equipment and medium. Background Art
[0002] In recent years, competition in the auto finance market has become increasingly fierce, and auto financing credit institutions have launched all-round competition in product innovation, risk control, and customer service.
[0003] Existing financing credit relies heavily on manual review for customer risk prevention and control, which makes it difficult to form a comprehensive and systematic risk assessment and fail to accurately judge a customer's potential credit risk. This results in low approval efficiency for auto finance credit applications and a deviation between credit risk and approval amount.
[0004] The patent document with application number 202010229807.8 discloses a method, device and medium for constructing a financial credit risk identification model. The scheme of this application can predict financial credit risks based on real-time financial events, enrich the dimensions of risk identification, and improve the predictability of credit risks of credit subjects. However, it also has the following problems: 1. It cannot be fully adapted to the field of automotive financial credit, and the credit investigation of customers cannot be quantified by AI automatic scoring. There is often a difference between the credit risk of customers obtained and the approved amount; 2. It is difficult to accurately judge the potential credit risk of customers, and there are problems such as low efficiency of financial credit approval.
[0005] The patent document with application number 202110214450.0 discloses an intelligent matching system for automobile financial products, which relates to the field of automobile financial technology. The application scheme tests the effects of the models generated by various artificial intelligence algorithms on the test set, selects the best-evaluated artificial intelligence algorithm and the corresponding matching discrimination model as the intelligent matching model of the automobile financial product, and generates the best intelligent matching discrimination model for each automobile financial product for all automobile credit products. At the same time, through the cooperation of blockchain and cloud storage, it stores the summary of customer personal information and financial information data on the blockchain, and stores the specific content of customer data on the cloud, thus realizing the complete protection of customer personal data. However, its scheme also has the following problems: its scheme focuses on the collection of customer credit qualifications, cannot conduct a comprehensive analysis of customer credit, and cannot form an AI automatic score quantification for customer credit. Summary of the invention
[0006] In view of the above problems, the purpose of the present invention is to provide an automobile financial credit AI robot risk control method, device, equipment and medium to perform comprehensive AI intelligent analysis of the customer's credit risk and accurately judge the customer's potential credit risk.
[0007] Embodiments of the present invention provide an automobile financial credit AI robot risk control method, device, equipment and medium.
[0008] Aspect 1: A risk control method for an auto finance credit AI robot, comprising:
[0009] S1. Conduct multi-level credit inquiries based on the basic information of the customer, and conduct intelligent analysis of the strategy set for each level of credit inquiry;
[0010] S2. Based on the intelligent analysis results of the strategy set, obtain the customer star rating and associate it with the system average default rate;
[0011] S3, performing multi-level function set intelligent analysis based on customer star rating information and customer vehicle financing information;
[0012] S4. Obtain the risk value of the star-second strategy rule based on the intelligent analysis results of the multi-level function set;
[0013] S5, according to the risk value of the star second strategy rule, select the strategy mode and perform risk decision diversion processing.
[0014] Furthermore, the S1 comprises the steps of:
[0015] S11. Obtain the customer's name, phone number, ID and bank card information;
[0016] S12. Conduct a credit inquiry on Pudao or Baihang based on the customer information, and analyze the credit regulation strategy set to obtain intelligent credit regulation scores, and perform exclusive processing based on the credit regulation scores;
[0017] S13. Continue to query Bairong's big data for customers who have passed the exclusive treatment, and analyze Bairong's strong regulation strategy set to obtain Bairong's strong regulation intelligent score, and perform exclusive treatment based on the Bairong strong regulation score;
[0018] S14. Continue to query the big data of excellent scores for the customers who have passed the exclusive treatment, and analyze the excellent score and strong rule strategy set to obtain the excellent score and strong rule intelligent score, and perform exclusive treatment according to the excellent score and strong rule score.
[0019] Further, the S2 includes:
[0020] S21. Based on the customer's credit score, Bairong score and Youfen score, the customer star rating is divided into the following levels: R, one star S, one star, two stars S, two stars, three stars S, three stars, four stars and five stars.
[0021] S22. The star ratings are associated with the average post-loan default rates of customers of each star level, which are:
[0022] The R-level default rate is 100%; the one-star S-related average default rate is 2 times; the one-star S-related average default rate is 1.6 times; the two-star S-related average default rate is 1.3 times; the two-star S-related average default rate is 1.1 times; the three-star S-related average default rate is 0.6 times; the three-star S-related average default rate is 0.3 times; the four-star S-related average default rate is 0.1 times; the five-star default rate is 0%.
[0023] Furthermore, the S3 comprises the steps of:
[0024] S31. Obtain the customer's vehicle financing information, including the financing amount, financing ratio and financing period;
[0025] S32. Based on the customer star rating information and the customer vehicle financing information, analyze the customer's comprehensive rate, financing limit, and credit limit to risk financing amount ratio function set.
[0026] Furthermore, the star-by-second strategy rules in S4 include:
[0027] frequency rules, rules of evidence, rules of location, and rules of agency.
[0028] Furthermore, the strategy modes in S5 include weight mode, worst match mode and first match mode:
[0029] The weight mode is that each policy rule is assigned a certain weight score, the risk score is calculated based on the hits of all policy rules, and the risk decision is made according to the risk score range that matches the risk score;
[0030] The first matching mode is that multiple policy rules are executed in order of priority, and the risk decision of the first policy rule hit is used as the final risk decision;
[0031] The worst matching mode is to execute all policy rules and take the decision with the most serious risk as the final risk decision.
[0032] Furthermore, the risk decision in S5 includes approval, rejection and transfer to manual.
[0033] One processing method is: perform exclusive processing based on the risk value of the Star Second Points strategy rules; after being rejected, perform Star Second Points rejection reason function calculation to analyze the reasons for Star Second Points rejection; after being approved, perform Star Second Points transfer to manual strategy set analysis, perform exclusive processing based on the analysis results, terminate the process and output the results after being approved, and transfer to manual processing after being rejected.
[0034] Another processing method is: exclusive processing based on the risk value of the Star Second Points strategy rules; after being rejected, the Star Second Points rejection reason function is calculated to analyze the reasons for the rejection of the Star Second Points; after being approved, the partner’s Star Second Points transfer to manual strategy set is analyzed; after the transfer to manual, the Star Seconds risk warning function is calculated to provide risk warnings for manual processing.
[0035] Aspect 2: An automotive finance credit AI robot risk control device, comprising:
[0036] Tag management unit, for customer information management;
[0037] The system dashboard unit comprehensively displays the overall operation decision-making and approval status of the system;
[0038] The credit index unit conducts multi-level credit inquiries and analyzes the indicators of credit at all levels;
[0039] The decision engine unit performs intelligent analysis of strategy sets and function sets to obtain customer star ratings and star-second scoring strategies;
[0040] The application management unit manages the application submission based on the system star rating and star-second scoring strategy;
[0041] Connect to the central unit and configure the policy set and function set;
[0042] Intelligent analysis unit, analyzes the star rating and star second scoring strategy results.
[0043] A third aspect: An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method provided in the first aspect when executing the program.
[0044] A fourth aspect: A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method provided in the first aspect.
[0045] Beneficial effects of the present invention:
[0046] 1. The present invention configures the strategy set and function set, performs AI intelligent analysis and calculation on the customer's credit information and credit information, accurately judges the customer's potential credit risk, automatically classifies and diverts the results, improves the approval efficiency of financial credit applications, accurately judges the customer's risk level and type, achieves a second-level response to vehicle financing credit approval, completes the timely diversion of second approval, second rejection and transfer to manual, and helps to simultaneously improve the credit business volume and customer quality.
[0047] 2. The present invention can display the AI approval status report of each applicant in real time, optimizing the customer experience. At the same time, various pre-loan and post-loan market indicators are visualized to form real-time monitoring, which can help risk control to optimize the model device in time.
[0048] 3. The average default rate of the star rating association system of the present invention can realize the association between the customer's quantitative score and the default situation, which can better prevent the customer from having credit risks.
[0049] 4. The risk decision of the present invention includes approval, rejection and transfer, supports joint customized rules and decision-making processes with partners, forms a joint risk prevention control, and provides risk control management level. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] Figure 1 This is a flow chart of a risk control method of an automobile financial credit AI robot according to the present invention;
[0051] Figure 2 It is a schematic diagram of the basic information of the customers and the decision results of the present invention;
[0052] Figure 3 A schematic diagram of the decision-making process of the present invention;
[0053] Figure 4 Another schematic diagram of the decision-making process of the present invention;
[0054] Figure 5 It is a schematic diagram of the decision-making mode of the present invention;
[0055] Figure 6 It is a schematic diagram of the composition of the risk control model device of the present invention;
[0056] Figure 7 This is a schematic diagram of the pre-loan market interface of the present invention;
[0057] Figure 8 This is a schematic diagram of decision flow diversion of the present invention;
[0058] Fig. 9 This is a schematic diagram of the application management of the present invention;
[0059] Fig.10 This is a schematic diagram of the risk report of the present invention;
[0060] Fig.11 It is a schematic diagram of the structure of the electronic device of the present invention. DETAILED DESCRIPTION
[0061] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar symbols throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.
[0062] In recent years, competition in the auto finance market has become increasingly fierce, and auto financing credit institutions have launched all-round competition in product innovation, risk control and customer service. However, the existing financing credit relies heavily on manual review for customer risk prevention and control, which makes it difficult to form a comprehensive and systematic risk assessment and cannot accurately judge the customer's potential credit risk, resulting in low approval efficiency for auto finance credit applications and a deviation between credit risk and approval amount.
[0063] In view of the above problems, the present invention provides an automobile financial credit AI robot risk control method. Figure 1 A flow chart of an automotive finance credit AI robot risk control method provided by an embodiment of the present invention, the method comprising:
[0064] S1. Conduct multi-level credit inquiries based on the basic information of the customer, and conduct intelligent analysis of the strategy set for each level of credit inquiry.
[0065] like Figure 2 As shown, the basic customer information includes the customer's name, telephone number, ID, bank card and other information; this information can be provided by the customer or a partner associated with the customer, such as a used car dealer who helps the customer apply for credit.
[0066] Pudao Credit, People's Bank of China Credit, Bairong and Youfen are third-party services. The credit information is provided by other systems and the system accesses the third party through the credit interface to query the credit information.
[0067] Pudao Credit, People's Bank of China Credit, Bairong and Youfen have different focuses and charging standards. They can directly complement each other and jointly reflect the credit situation of customers in different aspects to form a credit report. The system analyzes the strategy set based on different credit reports to obtain the customer's star rating. For example, Figure 3 Or as shown in 4, the steps may include:
[0068] S11. Obtain the customer's name, phone number, ID and bank card information;
[0069] S12. Conduct a credit inquiry on Pudao or Baihang based on the customer information, and analyze the credit regulation strategy set to obtain intelligent credit regulation scores, and perform exclusive processing based on the credit regulation scores;
[0070] S13. Continue to query Bairong's big data for customers who have passed the exclusive treatment, and analyze Bairong's strong regulation strategy set to obtain Bairong's strong regulation intelligent score, and perform exclusive treatment based on the Bairong strong regulation score;
[0071] S14. Continue to query the big data of excellent scores for the customers who have passed the exclusive treatment, and analyze the excellent score and strong rule strategy set to obtain the excellent score and strong rule intelligent score, and perform exclusive treatment according to the excellent score and strong rule score.
[0072] Among them, the credit reporting mandatory regulation strategy set, Bairong mandatory regulation strategy set and Youfen mandatory regulation strategy set mainly include: loan overdue situation, credit card overdue situation, maximum loan overdue period, maximum credit card overdue period, special transaction types of credit cards, guarantor compensation, debt-equity swap, judicial recovery, asset divestiture, asset transfer, compulsory installment, guarantor repayment, compulsory early repayment, compulsory execution record, bad debts and judicial recovery strategies.
[0073] Different strategy sets are formed based on the combination of the above-mentioned individual strategies, so as to obtain different scores according to the customer's situation.
[0074] S2. Based on the intelligent analysis results of the strategy set, obtain the customer star rating and associate it with the system average default rate. Specifically.
[0075] S21. Based on the customer's credit score, Bairong score and Youfen score, the customer star rating is divided into the following levels: R, one star S, one star, two stars S, two stars, three stars S, three stars, four stars and five stars.
[0076] S22. The star rating is associated with the average default rate of customers at each star level after loan issuance. The default rates of customers at each star level after association are:
[0077] The R-level default rate is 100%; the one-star S-related average default rate is 2 times; the one-star S-related average default rate is 1.6 times; the two-star S-related average default rate is 1.3 times; the two-star S-related average default rate is 1.1 times; the three-star S-related average default rate is 0.6 times; the three-star S-related average default rate is 0.3 times; the four-star S-related average default rate is 0.1 times; the five-star default rate is 0%.
[0078] Through the above-mentioned star-level classification of customers, the customer's credit scores from multiple sources can be integrated and risk-correlated with the average default rate to prepare for credit analysis.
[0079] S3. Perform intelligent analysis of multi-level function sets based on customer star rating information and customer vehicle financing information.
[0080] Customer vehicle financing information mainly includes information such as financing amount, financing ratio and financing period; this information can be obtained by partners associated with the customer. For example, a customer needs to buy a used car, the total amount required is 500,000, the customer's intended financing amount is 200,000, the financing ratio is 20 / 50=40%, and the financing period is 5 years.
[0081] Based on this vehicle financing information and combined with the customer's star rating information, we calculate and analyze the customer's comprehensive rate function, financing limit function, and credit limit and risk financing amount ratio function set.
[0082] S4. Obtain the risk value of the star-by-second strategy rule based on the intelligent analysis results of the multi-level function set.
[0083] Xingmiaofen comprehensively judges the vehicle financing information when the customer applies for the vehicle financing, and Xingmiaofen makes a decision in the automatic approval process, such as Figure 5 As shown, the Star Seconds strategy rules include: frequency rules, evidence rules, location rules and agency rules, etc., and each rule has its own score.
[0084] S5, according to the risk value of the star second strategy rule, select the strategy mode and perform risk decision diversion processing.
[0085] like Figure 5 As shown, the strategy modes may include weight mode, worst match mode and first match mode, and the risk decisions include approval, rejection and manual transfer.
[0086] Xingmiaofen mainly makes decisions and judgments in the automatic approval stage. The three risk decisions are approval (instant approval), rejection (instant rejection) and transfer to manual approval. Instant approval is for high-quality customers with high qualifications, and the process goes directly to the loan disbursement stage after making a judgment; instant rejection is for customers with poor qualifications, and a direct rejection is made; transfer to manual approval is for customers between instant approval and instant rejection. These customers may have information that cannot be captured by big data and need to be reviewed manually. After the judgment in the process, they will jump to the company's manual approval department.
[0087] In the weighted mode, each policy rule is configured with a certain weight score, and the risk score is calculated based on the hits of all policy rules. Risk decisions are made according to the risk score range that matches the risk score. When the policy mode is the weighted mode, the risk decision corresponding to the risk score needs to be configured. As shown in Table 1,
[0088] Table 1 Risk score allocation table
[0089] Score range Risk Decision 0~600 Pass (approved in seconds) 600~800 Transfer to manual 800~1000 Reject (reject immediately)
[0090] The first matching mode is that multiple policy rules are executed in sequence according to their priority, and the risk decision of the first matching policy rule is used as the final risk decision; the first matching mode has a sorting function.
[0091] The worst matching mode is to execute all policy rules and take the decision with the most serious risk as the final risk decision.
[0092] There are two main approaches to risk decision making:
[0093] like Figure 3 As shown, one processing method is: exclusive processing is performed according to the risk value of the star second point strategy rule; after being rejected, the star second point rejection reason function is calculated to analyze the reason for the star second point rejection; after being passed, the star second point transfer to manual strategy set analysis is performed, and exclusive processing is performed according to the analysis results. After being passed, the process is terminated and the result is output. After being rejected, it is transferred to manual processing.
[0094] This method uses the system to process by itself, and can be processed in three ways: approval (seconds approval), rejection (seconds rejection) and manual transfer. Each method will output a risk assessment report, such as Fig.10 As shown, Fig.10 Risk assessment report for rejection.
[0095] like Figure 4 As shown, another processing method is: exclusive processing based on the risk value of the Star Second Points strategy rule; after being rejected, the Star Second Points rejection reason function is calculated to analyze the reasons for the rejection of the Star Second Points; after being approved, the Star Second Points transfer to manual strategy set of the partner is analyzed; after the transfer to manual, the Star Seconds risk warning function is calculated to provide risk warnings for manual processing.
[0096] This method uses the system to conduct joint approval with partners. Partners have independent approval rights, and both those transferred manually and those passed can be submitted to partners for review.
[0097] The present invention also discloses an automobile financial credit AI robot risk control device for the above method, such as Figure 6 As shown, the device comprises:
[0098] The label management unit can be used to manage and maintain customer information and partner information.
[0099] System inventory unit, including pre-loan inventory, post-loan inventory and verification inventory, etc. Figure 7As shown, through the pre-loan dashboard interface, you can have an overview of the data and view the number of applications received under all decision results in the current system, the total application amount, the cumulative credit amount approved, the average credit amount, the average credit interest rate, the approval status, the hit rules, the correlation analysis, the trend table, the third-party call and the customer group analysis, etc., to comprehensively display the overall operation decision and approval status of the system.
[0100] The credit index unit conducts multi-level credit inquiries and analyzes the indicators of each level of credit. The credit index unit can parse the fields of the credit report formed by inquiries at all levels, obtain key characters, and allow the system to recognize the content of the credit report.
[0101] The decision engine unit performs intelligent analysis of strategy sets and function sets to obtain customer star ratings and star-second scoring strategies;
[0102] The decision engine unit mainly uses the policy set and function set configured by the model device to perform intelligent analysis, including the scoring rule management of various policies in the model device, as well as the decision flow management and diversion. Rule management is generally used to implement the model device parameter score processing and diversion rule set (transfer to manual rule set, rejection rule set, etc.). Decision flow management is used to complete the configuration of the entire model device. It can not only access third-party data sources, but also realize the decision-making function of this system (second approval, second rejection, transfer to manual), and can also refine management and empower different partners by configuring different decision flows, such as Figure 8 shown.
[0103] The application management unit manages the system's star-level and star-second strategy applications. Applications mainly include star-level applications and star-second strategy applications. Application management can present AI approval reports of customer applications in real time, as well as batch export of data for post-loan analysis, such as Fig. 9 As shown, customer submissions are displayed in a list in this module.
[0104] Connect to the central unit, configure policy sets and function sets, and through field management and configuration, form different policy or function sets to assist the decision engine in making decisions.
[0105] The intelligent analysis unit analyzes the star rating and star-second scoring strategy results, which can help analyze the trigger quantity and trigger results of various indicators of the model device. For example, it can analyze the reasons for passing (second approval), rejecting (second rejection) or transferring to manual processing.
[0106] The model device used in the present invention has an algorithm in the decision-making process based on a Bayesian classifier and a neural network. The Bayesian classifier is a probabilistic classification model device based on the Bayesian theorem and the feature conditional independence hypothesis, which realizes data classification by calculating the posterior probability, and its calculation formula is as follows:
[0107]
[0108] Among them: P(A|B) is the probability of event A occurring under the condition that event B occurs, called the posterior probability; P(B|A) is the probability of event B occurring under the condition that event A occurs; P(A) is the prior probability of event A occurring, that is, the probability of event A occurring in the absence of any other information; P(B) is the marginal probability of event B occurring, that is, the probability of event B occurring under any conditions.
[0109] The advantage of the Bayesian classifier is that it can integrate prior knowledge, which is particularly useful for situations where data is scarce, and by using prior knowledge, the Bayesian method can reduce the risk of overfitting. At the same time, the algorithm can handle various types of data, including discrete and continuous data, and the Bayesian classifier can be used not only for binary classification problems, but also for multi-classification problems.
[0110] In order to address the drawbacks of the Bayesian classifier algorithm, such as the possible inappropriateness of prior knowledge determination and data sparsity problems, the present invention combines the neural network algorithm of a deep learning model device (a computational model device that simulates the structure and function of the human brain, composed of a large number of nodes and their connections, and adjusts weights through learning to solve complex problems such as classification, recognition and prediction), optimizes the analysis and generalization capabilities of the model device, and improves the ability to process unstructured data and the scalability of the model device.
[0111] The model device of the present invention combines the advantages of Bayesian classifiers and neural networks in machine learning algorithms, optimizes the problems of overfitting, underfitting, and slow model operation speed in the processing of data by mainstream model algorithms in the automobile finance market, has strong analytical and generalization capabilities, can accurately capture data features, and has high operating speed and scalability. In terms of algorithm, it can fully identify customer characteristics, help the model device to perform accurate analysis, and realize precise and rational risk control. In terms of implementation, it achieves a response in seconds, greatly improving the approval efficiency of automobile financial credit applications.
[0112] The present invention also provides an electronic device, Fig.11 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Fig.11 As shown, the electronic device may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus. The processor may call the logic instructions in the memory, for example, to execute the following method:
[0113] S1. Conduct multi-level credit inquiries based on the basic information of the customer, and conduct intelligent analysis of the strategy set for each level of credit inquiry;
[0114] S2. Based on the intelligent analysis results of the strategy set, obtain the customer star rating and associate it with the system average default rate;
[0115] S3, performing multi-level function set intelligent analysis based on customer star rating information and customer vehicle financing information;
[0116] S4. Obtain the risk value of the star-second strategy rule based on the intelligent analysis results of the multi-level function set;
[0117] S5, according to the risk value of the star second strategy rule, select the strategy mode and perform risk decision diversion processing.
[0118] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention 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, including several instructions to enable 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 method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.
[0119] An embodiment of the present invention further provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method provided in each of the above embodiments is implemented, for example, including:
[0120] S1. Conduct multi-level credit inquiries based on the basic information of the customer, and conduct intelligent analysis of the strategy set for each level of credit inquiry;
[0121] S2. Based on the intelligent analysis results of the strategy set, obtain the customer star rating and associate it with the system average default rate;
[0122] S3, performing multi-level function set intelligent analysis based on customer star rating information and customer vehicle financing information;
[0123] S4. Obtain the risk value of the star-second strategy rule based on the intelligent analysis results of the multi-level function set;
[0124] S5, according to the risk value of the star second strategy rule, select the strategy mode and perform risk decision diversion processing.
[0125] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.
[0126] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. An automobile finance credit AI robot risk control method, characterized in that: include: S1. Conduct multi-level credit inquiries based on the basic information of the customer, and conduct intelligent analysis of the strategy set for each level of credit inquiry; S2. Based on the intelligent analysis results of the strategy set, obtain the customer star rating and associate it with the system average default rate; S3, performing multi-level function set intelligent analysis based on customer star rating information and customer vehicle financing information; S4. Obtain the risk value of the star-second strategy rule based on the intelligent analysis results of the multi-level function set; S5, according to the risk value of the star second strategy rule, select the strategy mode and perform risk decision diversion processing.
2. The automobile finance credit AI robot risk control method according to claim 1 is characterized in that: The S1 comprises the steps of: S11. Obtain customer name, phone number, ID and bank card information; S12. Conduct a credit inquiry on Pudao or Baihang based on the customer information, and analyze the credit regulation policy set to obtain intelligent credit regulation scores, and perform exclusive processing based on the credit regulation scores; S13. Continue to query Bairong's big data for customers who have passed the exclusive treatment, and analyze Bairong's strong regulation strategy set to obtain Bairong's strong regulation intelligent score, and perform exclusive treatment based on the Bairong strong regulation score; S14. Continue to query the excellent score big data for the customers who have passed the exclusive treatment, and analyze the excellent score and strong rule strategy set to obtain the excellent score and strong rule intelligent score, and perform exclusive treatment according to the excellent score and strong rule score.
3. The automobile finance credit AI robot risk control method according to claim 1 is characterized in that: The S2 includes: S21. Based on the customer's credit score, Bairong score and Excellent score, the customer star rating is divided into: R, one star S, one star, two stars S, two stars, three stars S, three stars, four stars and five stars; S22. The star ratings are related to the average post-loan default rates of customers of each star level, which are: The R-level default rate is 100%; the one-star S-related average default rate is 2 times; the one-star S-related average default rate is 1.6 times; the two-star S-related average default rate is 1.3 times; the two-star S-related average default rate is 1.1 times; the three-star S-related average default rate is 0.6 times; the three-star S-related average default rate is 0.3 times; the four-star S-related average default rate is 0.1 times; the five-star default rate is 0%.
4. The automobile finance credit AI robot risk control method according to claim 1 is characterized in that: The S3 comprises the steps of: S31. Obtain the customer's vehicle financing information, including the financing amount, financing ratio and financing period; S32. Based on the customer star rating information and the customer vehicle financing information, analyze the customer's comprehensive rate, financing limit, and credit limit to risk financing amount ratio function set.
5. The automobile finance credit AI robot risk control method according to claim 1, characterized in that: The star-second strategy rules in S4 include: frequency rules, evidence rules, location rules and agency rules.
6. The automobile finance credit AI robot risk control method according to claim 1 is characterized in that: The strategy modes in S5 include weight mode, worst match mode and first match mode, where: The weight mode is that each policy rule is assigned a certain weight score, and the risk score is calculated based on the hits of all policy rules. The risk decision is made according to the risk score range that matches the risk score. The first matching mode is that multiple policy rules are executed in order of priority, and the risk decision of the first policy rule hit is used as the final risk decision; The worst matching mode is to execute all policy rules and take the decision with the most serious risk as the final risk decision.
7. The automobile finance credit AI robot risk control method according to claim 1 is characterized in that: The risk decisions in S5 include approval, rejection and manual transfer: One processing method is: exclusive processing is performed according to the risk value of the Star Second Points strategy rule; after being rejected, the Star Second Points rejection reason function is calculated to analyze the reasons for the Star Second Points rejection; after being approved, the Star Second Points transfer to manual strategy set analysis is performed, and exclusive processing is performed according to the analysis results. After being approved, the process is terminated and the results are output. After being rejected, the process is transferred to manual processing; Another way to handle it is to perform exclusive processing based on the risk value of the Star Second Points strategy rules; after being rejected, calculate the Star Second Points rejection reason function and analyze the reasons for the Star Second Points rejection; after being approved, analyze the Star Second Points transfer to the manual strategy set of the partner; After the transfer to manual work, the Star Second risk warning function calculation is performed to provide risk warnings for manual processing.
8. An AI robot risk control device for automobile financial credit applied to the method according to any one of claims 1 to 7, characterized in that: The device comprises: Tag management unit, for customer information management; The system dashboard unit comprehensively displays the overall operation decision-making and approval status of the system; The credit index unit conducts multi-level credit inquiries and analyzes the indicators of credit at all levels; The decision engine unit performs intelligent analysis of strategy sets and function sets to obtain customer star ratings and star-second scoring strategies; The application management unit manages the application submission based on the system star rating and star-second scoring strategy; Connect to the central unit and configure the policy set and function set; Intelligent analysis unit, analyzes the star rating and star second scoring strategy results.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of an automobile financial credit AI robot risk control method as described in any one of claims 1 to 7 are implemented.
10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of an automobile financial credit AI robot risk control method as described in any one of claims 1 to 7 are implemented.
Citation Information
Patent Citations
Financial credit risk identification methods, model building methods and devices
CN111383102B
Intelligent matching system for automobile financial products
CN112907354A
Borrowing and lending business decision-making method and system based on a business rule base
CN109493213A
Automobile financial post-loan risk early warning method
CN113298394A
Multi-level evaluation method, device and equipment for credit customer and storage medium
CN115271934A
