Low-code client illegal fund transfer risk rating method and system

Abstract

The invention relates to a low-code client illegal fund transfer risk rating method and system, and the method comprises the steps: receiving data when a client transactions through a unified rating interface, comparing the data with preset risk elements, carrying out the score accumulation through the division and weighting of risk dimensions and risk subclasses set for the risk elements in advance, and carrying out the calculation of the risk factors. Obtaining an illegal fund transfer risk rating result; the rating configuration of the illegal fund transfer risk is displayed and edited through a front-end visual interface; a precondition and a postcondition are set for the illegal fund transfer risk rating, and if the received data meet a matching mode of the precondition, a score of the illegal fund transfer risk rating is set; and if the received data meets the post condition, adjusting the final rating score or risk grade. Compared with the prior art, the method has the advantages that the risk rating process is more flexible and efficient, and the rating consideration range is wider.

Description

Technical Field

[0001] The present invention relates to the field of big data service technology, and in particular to a low-code customer illegal funds transfer risk rating method and system. Background Art

[0002] In the framework of the anti-illegal fund transfer and anti-terrorist financing compliance management system of financial institutions, customer risk rating is the cornerstone of the compliance management system. Therefore, it is particularly important to establish a visually configurable, programmable and extensible risk-based method for identifying customer illegal fund transfer / terrorist financing risk rating.

[0003] Most customer risk rating methods start with basic customer information, geographic location information, transaction activities, and industry information, and draw up a detailed list of rating indicators to calculate the customer's risk level. This scorecard model indicator system is relatively mature. However, when faced with changes in different regulatory regions and different regulatory preferences, customized development of business processes is required.

[0004] In order to cope with the increasingly severe international and domestic situation against illegal funds transfer, meet global financial regulatory compliance requirements, and reduce development pressure, there is an urgent need for a "low-code" customer illegal funds transfer risk rating calculation method. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a low-code customer illegal funds transfer risk rating method and system.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] A low-code customer illegal funds transfer risk rating method, comprising receiving customer transaction data through a unified rating interface, and comparing it with pre-set risk factors to perform illegal funds transfer risk rating;

[0008] The rating configuration of the risk of illegal fund transfer is displayed and edited through a front-end visual interface, wherein the rating configuration of the risk of illegal fund transfer includes the weight and correlation strength of the risk factors.

[0009] Furthermore, the illegal funds transfer risk rating is provided with a precondition, and the precondition is provided with a precondition matching mode. If the received data satisfies the precondition matching mode, the score of the illegal funds transfer risk rating is set.

[0010] Furthermore, when the precondition matching mode is directed to a specific risk level, if the data satisfies the precondition matching mode, the final risk rating score is directly set to the preset first score.

[0011] Further, when the precondition matching pattern is directed to a specific risk level and above, if the data satisfies the precondition matching pattern, the final risk rating score is set to at least a preset second score.

[0012] Furthermore, the illegal funds transfer risk rating is provided with a post-condition, and if the received data satisfies the post-condition, the final rating score or risk level is adjusted.

[0013] Furthermore, the rating process of the risk of illegal fund transfer includes:

[0014] Multiple risk dimensions are preset, each risk dimension is provided with multiple risk subcategories, each risk subcategories corresponds to one of the risk factors, the risk subcategories under each risk dimension are provided with corresponding subcategories weights, and the sum of the subcategories weights under the risk dimension is 1, each risk dimension is provided with corresponding dimension weights, and the sum of all dimension weights is 1;

[0015] Data matching is performed through the risk factors corresponding to each risk subcategory to obtain the score of each risk subcategory, and then the score of each risk dimension is obtained by multiplying it by the corresponding subcategory weight and accumulating it. Finally, the risk score is obtained by multiplying it by the final dimension weight and accumulating it.

[0016] Furthermore, the subclass weights of the risk subclasses are allocated based on the corresponding association strengths.

[0017] Furthermore, each risk factor is scored using a binary scoring method, a relative scoring method, or an absolute scoring method.

[0018] Furthermore, the front-end visualization interface is also used to add or delete risk factors;

[0019] When adding a risk factor, first determine the risk dimension to which the risk factor belongs, then select the corresponding scoring method for the risk factor, and finally assign a weight to the risk factor;

[0020] When a risk factor is deleted, the subclass weight corresponding to the risk factor is set to 0.

[0021] The present invention also provides a low-code customer illegal funds transfer risk rating system, including a memory and a processor, wherein the memory stores a computer program, and the processor calls the computer program to execute the steps of the method described above.

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

[0023] (1) The risk customer rating method of the present invention is a "low-code" and flexibly configurable solution. On the one hand, business personnel do not need to write code. They can see the display of these configurations on the front end, and can add, edit, and deactivate preconditions and postconditions. They can also edit and adjust the weights and correlation strengths of risk factors according to different regulatory requirements. On the other hand, for developers, when they need to add or modify risk factors and preconditions, it is also "low-code". They only need to write a small amount of code, or even do not need to rewrite the code. They only need to complete the update of the parameter configuration table. Such a "low-code" design makes risk rating more flexible and efficient.

[0024] (2) The present invention pioneered the concept of preconditions and postconditions. The addition of preconditions and postconditions makes the scope of customer risk rating more extensive and the rating more standardized. Preconditions make the rating more focused and efficient. For example, the hit list is an important point in identifying illegal fund transfers and terrorist financing activities. Therefore, it is first necessary to identify whether the customer is on the hit list. If so, the customer is assigned to the corresponding risk level or above. Postconditions serve as auxiliary risk factors and as a supplement. Both have an important impact on customer risk ratings.

[0025] (3) When the present invention is used to calculate customer risk scores, parameters can be flexibly configured, which can meet the needs of different institutions for different supervision. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A flowchart of a low-code method for rating the risk of illegal funds transfer by customers provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0028] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0030] It should be noted that the relevant terms in this embodiment are explained as follows:

[0031] Low-code: Through visualization (graphical user interface, using drag-and-drop components and rule-driven processes to create or modify program logic), or through a small amount of code, applications can be quickly generated, so that developers with different experience levels or even business personnel without coding foundation can quickly develop applications or modify business logic. The "low-code" in the present invention is essentially a process of summarizing, abstracting and templating the application scenarios of customer illegal funds transfer risk rating.

[0032] Risk factors: The quantification of a client’s risk of illegal fund transfer and terrorist financing should focus on all factors related to the client, such as the client’s background information, occupation, source of funds or income, client geographic information (client identity, location, place of transaction, main business area, etc.), products and services used by the client, the nature and purpose of the account opened by the client, the expected or actual use of the account (the type, volume, amount and frequency of transactions conducted by the client using the account), associated accounts, business activities and whether there are significant changes in the bank-client relationship.

[0033] Binary scoring method: This scoring method is applicable when the question information is binary in nature and the answer to the question can be classified as "yes / no". If the customer has the risk characteristics of a certain risk factor, he should get 5 points; if the customer does not have the risk characteristics of the risk factor, he should get 1 point.

[0034] Relative scoring method: This method compares the quantitative data corresponding to all customers, sorts them by size, and divides them into N intervals (N=5 in this method). The score of each customer depends on which interval their quantitative data falls into. This scoring method is mainly used for risk factors driven by quantitative data (for example, for the risk factor of "cash inflow", if the transaction amount of a certain individual customer during the review period is in the period [0,2k), [3w,10w), [10w,+∞) respectively, then it should be scored 1, 3, and 5 points respectively). For different types of customer groups, the corresponding intervals and scores will also be adjusted accordingly.

[0035] The use of relative scoring method enables risk thresholds to be generated based on the analysis of the actual distribution of bank customers, so that risk scores are more in line with the unique customer base risk situation of different customer groups.

[0036] Absolute scoring: This scoring method is suitable for scenarios where the risk factor cannot be scored relative to the risk factor or where it is necessary to identify high-risk characteristics specified in regulatory requirements and industry guidelines (for example, for the risk factor "industry", if the individual customer's industry is precious metals and jewelers, it should be scored 5 points; if it is higher education, it should be scored 1 point). Absolute scoring is mainly used for risk factors with non-numeric values, such as industry, occupation, region, or type of corporate document / identification.

[0037] Score level mapping: After the calculator calculates the customer's risk score based on the configured factors, it will classify the customer into one of the following levels: A, B, C, D, or E as the final risk rating result (A is the lowest risk and E is the highest risk) based on the score level mapping rules.

[0038] In order to meet the local regulatory requirements of each overseas branch, the rules for mapping the score level interval can be flexibly configured. For example, if the risk scores are in the range of (-1,2], (2,2.5], (2.5,3.1], (3.1,3.5], (3.5,5], they should be recorded as A, B, C, D, E respectively.

[0039] Preconditions: "Preconditions" are the main risk factors and will be given priority. The results of such conditions will have a greater impact on the rating. For example, when a customer hits the sanctions list, the customer's risk level can be directly set to E level, and the rules for other risk factors will no longer be calculated. In addition, preconditions can also be specified above a specific risk level.

[0040] Post-condition: The post-condition is used as an auxiliary risk factor. If the customer meets the corresponding post-condition, the corresponding score will be increased (customer positive information) / reduced (customer negative information) based on the risk score that has been obtained, or the risk level will be raised to a higher level / lowered to a lower level based on the risk score that has been obtained.

[0041] Example 1

[0042] This embodiment provides a low-code customer illegal funds transfer risk rating method, including receiving customer transaction data through a unified rating interface, and comparing it with pre-set risk factors to perform illegal funds transfer risk rating;

[0043] The rating configuration of the risk of illegal fund transfer is displayed and edited through a front-end visual interface, wherein the rating configuration of the risk of illegal fund transfer includes the weight and correlation strength of the risk factors.

[0044] The IFT risk rating process mainly includes:

[0045] 1) A precondition, which is provided with a precondition matching mode. If the received data satisfies the precondition matching mode, the score of the illegal funds transfer risk rating is set.

[0046] When the precondition matching mode is directed to a specific risk level, if the data satisfies the precondition matching mode, the final risk rating score is directly set to the preset first score.

[0047] When the precondition matching mode is directed to a specific risk level or above, if the data satisfies the precondition matching mode, the final risk rating score is set to at least a preset second score.

[0048] 2) Main process:

[0049] Multiple risk dimensions are preset, each risk dimension is provided with multiple risk subcategories, each risk subcategories corresponds to one of the risk factors, the risk subcategories under each risk dimension are provided with corresponding subcategories weights, and the sum of the subcategories weights under the risk dimension is 1, each risk dimension is provided with corresponding dimension weights, and the sum of all dimension weights is 1;

[0050] Data matching is performed through the risk factors corresponding to each risk subcategory to obtain the score of each risk subcategory, and then the score of each risk dimension is obtained by multiplying it by the corresponding subcategory weight and accumulating it. Finally, the risk score is obtained by multiplying it by the final dimension weight and accumulating it.

[0051] The sub-class weights of each risk sub-class are allocated based on the corresponding strength of association. Each risk factor is scored using a binary scoring method, a relative scoring method, or an absolute scoring method.

[0052] 3) Post-condition: If the received data meets the post-condition, the final rating score or risk level is adjusted.

[0053] The front-end visual interface is also used to add or delete risk factors;

[0054] When adding a risk factor, first determine the risk dimension to which the risk factor belongs, then select the corresponding scoring method for the risk factor, and finally assign a weight to the risk factor;

[0055] When a risk factor is deleted, the subclass weight corresponding to the risk factor is set to 0.

[0056] This embodiment also provides a low-code customer illegal funds transfer risk rating system, including a memory and a processor, the memory storing a computer program, and the processor calling the computer program to execute the steps of the low-code customer illegal funds transfer risk rating method as described above.

[0057] like Figure 1 As shown, the implementation process of the low-code customer illegal funds transfer risk rating method of this embodiment is described in detail below.

[0058] 1. Unified rating interface standards

[0059] Provide a bank-wide rating interface for other systems for real-time / batch rating of customers' illegal fund transfer risks. The interface provides unified rating content and format, including subject information, basic transaction information, etc. The upstream system sends the information to the system through message middleware such as WebSphere MQ or Kafka, and completes customer real-time risk rating, account opening transactions, other transactions, etc. through distributed processing, and then returns the rating and transaction processing results through the message middleware.

[0060] 2. Visualized risk factor parameters, risk subcategories and weight configurations

[0061] This section explains the configuration of risk factor parameters through the "relative scoring method" and "absolute scoring method":

[0062] Based on the huge total number of customers, financial institutions need to implement a risk-based management philosophy, consider the rationality of the number of high-risk customers and the required management resources, control measures, etc. The risk-based approach requires financial institutions to set different risk level target distributions for different customer types.

[0063] Table 1

[0064]

[0065]

[0066] As can be seen from Table 1, among all natural person customers of financial institutions under the jurisdiction, those with medium to high risk or above account for 7%. The use of the relative scoring method can generate risk thresholds based on the analysis of the actual distribution of bank customers, so that the risk score is more in line with the unique customer base risk situation of financial institutions. Take the configuration adjustment of the risk factor "cash outflow" as an example, as shown in Table 2:

[0067] Table 2

[0068]

[0069] With the continuous improvement of economic development and people's income and expenditure levels, the disposable income of residents in different countries or different provinces / states in the same country also varies. Therefore, different institutions have different requirements for the interval and score configuration of this risk factor. Business personnel can directly modify it through the front-end visual interface and flexibly adjust the upper and lower limits of the transaction amount corresponding to the risk score. For example, there may be another scoring interval for the "cash outflow" risk factor, as shown in Table 3:

[0070] Table 3

[0071]

[0072] This method is applicable to parameter adjustments based on transaction risk factors (i.e. cash inflow, wire transfer inflow, wire transfer outflow, etc.). By flexibly adjusting the parameter range, the model can be adjusted in the next second without stopping the machine, and the high adaptability of the entire method is established instantly!

[0073] Another example: the customer's age. Age will reflect some tangible or intangible characteristics of the customer, such as economic ability, risk preference, ability to accumulate legal income, etc. Based on the bank's existing historical data, the age risk factor is scored according to the proportion of customers in each age range who are the main subjects of suspicious transaction reports, and the score range for the "customer age" risk factor is given as shown in Table 4.

[0074] Table 4

[0075]

[0076] As the social demographic structure changes or due to unbalanced regional economic development, different countries or different provinces / states in the same country can flexibly adjust the age range corresponding to the risk score based on the ratio of the population of the administrative division to the customer data of the subject of the suspicious transaction report. There is another scoring standard for the "customer age" risk factor, as shown in Table 5:

[0077] Table 5

[0078]

[0079] Except for "customer age", parameter adjustments for most risk factors that take non-numeric values ​​(industry, occupation, region, or type of corporate document / identification) are applicable to the absolute scoring method.

[0080] Note some exceptions: In the case of default scoring, when a customer is missing the required data for a risk factor, this method will assign the customer the average score of all customer categories (i.e., natural persons, enterprises, or financial institutions) for that risk factor. For example, when a natural person customer is missing monthly income data, the customer's monthly income risk factor score is the average of all natural person customer monthly income risk factor scores.

[0081] Of course, the default score is also configurable according to the organization's situation. Business personnel can assign scores to missing values ​​on their own. For example, if a natural person customer is missing important risk factor data, the score for that factor can be directly set to 5 points.

[0082] A client's IFT / Terrorist Financing risk is a composite score comprised of four main risk subcategories: client characteristics, client geography, client product and service usage (including transaction activity), and client industry / occupation.

[0083] Each risk sub-category is calculated using the "weight method", and the sum of the weights is 100% of the customer risk score under that dimension. The four risk dimensions also use the "weight method" to calculate the final customer risk score, and the sum of the weights of each risk dimension is 100% of the customer's final risk score. The weight of each risk dimension depends on the allocation of the strength of correlation between the risk classification and the risk of illegal fund transfer by each institution. The strength of correlation of risk classification is distinguished from low to high with a score of "1-5", and the weight of a risk classification refers to the percentage of its correlation strength in the combined strength of correlation of all risk classifications, as shown in Table 6:

[0084] Table 6

[0085]

[0086]

[0087] The calculation rules for each risk sub-category under a risk dimension are similar. The weight of the risk sub-category depends on the distribution of its correlation strength (the correlation strength is also divided from low to high with a score of "1-5"). Taking the "customer region" dimension as an example, the specific calculation rules are shown in Table 7:

[0088] Table 7

[0089]

[0090]

[0091] (Note: The strength of the correlation between each risk sub-factor depends on the determination of the strength of the correlation between the risk factor and the risk of illegal fund transfer by each institution based on regulatory requirements and existing documents. Therefore, the weight configuration of the calculators of different institutions may vary.)

[0092] The purpose of using numerical ratios to calculate the weights of risk subclasses and risk factors is to simplify the process of adding or deleting risk subclasses and risk factors. We no longer need to manually adjust the weights for each risk subclass and risk factor. By using the above method, the weight adjustment of any risk subclass will adjust the weight values ​​of other risk subclasses in a linked manner. Similarly, under the same risk subclass, the weight adjustment of any risk factor will adjust the weight values ​​of other risk factors of the same type in a linked manner. As shown in Table 8:

[0093] Table 8

[0094]

[0095]

[0096] Taking financial institutions as an example, if annual income needs to be considered as a risk factor one day and its correlation strength is set to 1, then after updating the database, the calculator will read the new weight configuration: the type of enterprise, duration of customer relationship, type of enterprise certificate, and annual income have corresponding weights of approximately 43%, 29%, 14%, and 14%, respectively.

[0097] When adding a new risk factor, you must first determine the nature of the new risk factor and confirm the risk dimension to which it belongs, and then evaluate the data availability and data quality of the new risk factor. And select one of the three scoring methods (binary scoring method, relative scoring method, absolute scoring method) for the new factor, and the risk score assigned to the new risk factor must be between 1 and 5 points. After the scoring method is determined, the weight of the factor is assigned. After completing the above steps, the new factor is added to the calculation of the customer risk level rating. Conversely, if you want to delete the existing risk factor that has been included in the model calculation, directly adjust the weight of this risk factor to 0 (the correlation strength is 0), and the subsequent customer risk rating score will no longer include it in the comprehensive score. The addition / discarding of risk factor items also reflects the flexible configurability of the method from one aspect. Different institutions, different customer groups, and different customer types can configure different risk factors, risk factor weights, and correlation strengths of each factor, so as to meet the compliance needs of different institutions.

[0098] This embodiment divides customer risk levels into five dimensions (five-level risk level classification):

[0099] Risk Level E – “High Risk”

[0100] Risk level D – “medium to high risk”

[0101] Risk Level C – “Medium Risk”

[0102] Risk Level B – “Low to Medium Risk”

[0103] Risk Level A – “Low Risk”

[0104] The present invention uses three independent customer risk rating models to assess customer risk:

[0105] (1) natural persons, (2) enterprises, (3) financial institutions;

[0106] At the same time, the applicable customer groups will be distinguished:

[0107] (1) existing customers, (2) new customers;

[0108] The three independent models in the present invention have the same calculation method, and the risk factor combinations used by different models are slightly different. The risk factors, factor weights, and score cutoff points used to divide risk levels (score level mapping) used by different models are all flexible and configurable. By providing a standardized online / batch service interface and adopting a pluggable strategy, real-time / batch customer rating is performed.

[0109] This embodiment identifies the illegal fund transfer and terrorist financing risks of customers from the following four risk subcategories:

[0110] 1)Customer characteristics;

[0111] 2) Customer region;

[0112] 3) Use of customer products and services (including trading activities);

[0113] 4) The client’s industry / occupation.

[0114] Each risk classification is composed of multiple risk sub-factors. For example, "customer characteristics" include basic customer information such as customer age, income, duration of customer relationship, account purpose, etc. In this method, we use "correlation strength" and "weight value" to measure the impact of these four dimensions and different risk sub-factors under the dimensions on the risk score. The stronger the correlation strength (the highest is 5 and the lowest is 1), the higher the weight of the risk factor. After the calculator reads the customer's real data, it takes the maximum value under this dimension or the weighted average method to obtain the customer's risk score under this dimension. In this method, each customer is scored from the four risk dimensions respectively, and the risk level score of the customer is obtained by weighted averaging the risk scores of the four dimensions.

[0115] 3. Visual pre- and post-condition configuration

[0116] The present invention introduces the concept of precondition and postcondition, which is first created in the industry. Because the calculator involves risk level mapping rules when executing preconditions and postconditions, a set of risk level mapping rules is first listed as a reference, and the rules are shown in Table 9.

[0117] Table 9

[0118] grade Score level mapping A (low risk) (-1,2] B (medium to low risk) (2,2.5] C (Medium risk) (2.5,3.1] D (Medium to High Risk) (3.1,3.5] E(High Risk) (3.5,5]

[0119] 3.1 Prerequisites

[0120] As the main risk factor, the prerequisites will be given priority when calling the calculator to calculate the risk level.

[0121] If the precondition matching mode is "pointing to a specific risk level" and is medium-high risk (risk level is D), if the customer hits this condition, his / her final risk rating score is directly set at 3.5 points. For customers who hit the specified risk level, there is no need to perform rule calculations for other factors.

[0122] If the matching mode of the precondition is "pointing to a specific risk level and above", it is specified as medium risk and above (risk level is C or above, i.e. C, D, E).

[0123] Customers who meet the precondition of "pointing to risk level and above" need to calculate the risk factors of four dimensions and the postconditions.

[0124] Since different institutions and different regulatory requirements focus on different points in risk rating, the preconditions of each institution are also different, and can be flexibly configured in the calculator of the present invention. Business personnel can see all the preconditions under the institution through the front end, and can add, edit, and deactivate preconditions. As described in the example of "identity hits high-risk countries" above, the matching mode and score of the condition can be adjusted through the front end.

[0125] Because it can directly point to a specific risk level, preconditions have a great impact on the customer risk rating results. From a functional perspective, after adding major risk factors (preconditions), there is no need to perform subsequent calculations for customers who hit the specified risk level preconditions, which can save calculation time and improve efficiency. From a rule configuration perspective, preconditions generally list relatively or obviously suspicious situations as hit conditions. If these preconditions are hit, there will be a basic judgment on the risk situation of the customer. Therefore, it is logical to give priority to preconditions.

[0126] In this embodiment, the preconditions include but are not limited to the following:

[0127] 3) Refusing to cooperate with customer due diligence;

[0128] 5) Has been the subject of a suspicious transaction report in the past five years;

[0129] 6) Other financial institutions identify them as high risk.

[0130] 7. …

[0131] Due to the different regulatory requirements of overseas branches, the present invention can configure different preconditions for different branches. For example, the branch can use some factors that are more important to the regulator as preconditions, that is, major risk factors, such as the industry of the customer, whether the customer is in a high tax risk country, etc. The risk rating calculator will automatically call different precondition rules when it is running.

[0132] 3.2 Postconditions

[0133] Postconditions serve as auxiliary risk factors and are conditions that are performed after the calculator has calculated the preconditions and the scores of the four-dimensional risk factors.

[0134] There are also two matching modes for postconditions. The first one is plus / minus items, such as Table 12:

[0135] Table 12

[0136] describe Matching Pattern Fraction Positive Information Deduction 0.1 Negative Information Bonus points 0.2

[0137] If a customer's score after calculation of the preconditions and four-dimensional risk factors is 3.2, and the "positive information" listed in the postconditions is hit, the final score is 3.1. According to the risk level mapping rules, the customer's final risk level is C.

[0138] The second is to increase / decrease the risk level, as shown in Table 13:

[0139] Table 13

[0140] describe Matching Pattern Fraction Positive Information Raise the risk - Negative Information Reduce risk -

[0141] At this time, if the customer's score after calculation based on the preconditions and four-dimensional risk factors is 3.2, according to the risk level mapping rules, its risk level is D. If the "negative information" in the postcondition is also met, the customer's final risk level is E.

[0142] Post-conditions include but are not limited to the following:

[0143] 1) Whether it is a listed company;

[0144] 2) Other financial institutions identify it as positive information;

[0145] 3) Due diligence by Chinese banks is a control measure against illegal fund transfers that needs to be improved;

[0146] 4) Lack of adequate control measures against illegal fund transfers;

[0147] 5) Other financial institutions identify it as negative information.

[0148] Post-conditions are also flexibly configurable. Business personnel can add or disable conditions through the front end. Similarly, they can also edit the matching mode and score of existing post-conditions.

[0149] Post-conditions can directly add or subtract fixed scores, or even raise or lower the risk level, so post-conditions are also one of the factors that have an important impact on the customer's risk level. In addition to considering whether the customer has obvious suspicious circumstances and basic information in four dimensions, the introduction of the concept of post-conditions makes the risk factors involved in the calculator more comprehensive. For example, some information that is not emphasized by the regulator, but that various institutions may find related to illegal fund transfers or terrorist financing can be listed as post-conditions. Comprehensive considerations can make the rules more complete.

[0150] 4. Rules Engine

[0151] In customer risk rating, the focus is on identifying the risk factor information when the customer conducts a transaction (establishing a business relationship or any business) and based on this data set, further assessing whether the customer is a high-risk customer for illegal fund transfer when the transaction occurs. The above-mentioned customer's account opening transactions, loan approval, foreign exchange purchase transactions, etc. initiate rating requests to this system in real time through online connection, so the use of a "pluggable" configuration mode meets the various requirements of this business scenario. The calculation mechanism of the present invention is as follows:

[0152] This method scores customer data using the scoring rules in "Risk subcategories and risk factors" and "Preconditions and postconditions" described in Sections 2 and 3 of this part. The steps for assessing the risk level of each customer are as follows:

[0153] The "unified rating interface standard" is the basis for the invention's external services. The "risk parameters, risk sub-categories and weight configuration" and "pre- and post-condition configuration" are the basis for the "rule engine". The two are dependent on each other. Both are the core factors for the stable and accurate operation of the rule engine.

[0154] The "flexibly configurable rule engine" is the application of "risk parameters, risk subcategories and weights" and "preconditions and postconditions" data. Only by using this method can the customer's illegal fund transfer / terrorist financing risks be extracted from the collection of global large data volumes, and calculated and fed back in a very short time.

[0155] The standardized interface provided by the present invention adopts an independent service mode, taking into account the seriousness of supervision, real-time nature of transactions, business access costs, etc., and provides compliance services for the risk rating of illegal fund transfers of customers for the entire trading system of the group. Furthermore, this universal design is also conducive to data interaction between institutions and expands the potential of real-time rating scope.

[0156] The customer risk rating methods used in the industry have shortcomings in risk factor selection, data collection, uncertainty of the model itself, and rating efficiency. This is reflected in the fact that the lack of certain data leads to model limitations, the imperfect data set based on it leads to the inaccuracy of the model, the high cost of integration and control of banks, insurance companies, securities companies, and fund companies operating independently, the lack of unified data exchange standards, and the poor effectiveness of ratings, especially when the number of customers at outlets surges, which affects the handling of other normal transactions. There are many different customer rating methods, and no advanced, global, and efficient rating method has yet to be formed.

[0157] This plan proposes a complete set of independent rating solutions to conduct real-time control over customers across the bank, and introduces the industry-first concept of pre-conditions and post-conditions, providing loosely coupled technical compliance services with almost no impact on the original business system processes.

[0158] In addition, most existing risk rating technologies use a scorecard-style indicator system, and a set of scoring standards may need to be redefined for different institutions and different regulatory needs. This solution, on the basis of proposing a complete set of independent risk rating solutions, increases the flexible configurability of the risk rating calculator. This "flexible configuration" is reflected in the fact that the relevant configuration of the calculator can be flexibly changed as regulatory requirements and business needs change. Business personnel do not need to write code, but can directly add, edit, and delete risk factors according to the help instructions on the front-end display page, and the new calculator configuration is effective immediately. On the other hand, developers can also quickly update applications by modifying a small amount of code, or even without rewriting the code, by simply modifying the parameters of the configuration table according to the new requirements. Such a "low-code" model not only reduces the pressure of development, but also improves the efficiency of system updates.

[0159] Based on the specific needs of each financial institution for preconditions, four-dimensional risk factors, and postconditions, the rating calculator used by each institution can be unique, which makes the customer's risk rating method more in line with local supervision and local regulations.

[0160] The present invention aims to give warnings when establishing business relationships with high-risk customers or handling any business with high-risk customers, and to improve business efficiency by providing universal interfaces to serve the entire bank system. The present invention achieves the above-mentioned goals.

[0161] It is particularly noted that the present invention requires different financial institutions to combine their own business characteristics and business scope to select risk factors and corresponding weights that match themselves and meet local regulatory requirements, but this method can be used as a general standard for financial institutions to conduct customer risk rating. Phase practice has shown that this method has achieved significant effectiveness. The same customer may have different risk levels in different financial institutions due to differences in the models used by the institutions or the different types and purposes of the financial institution's products used by the customer, which is not within the scope of consideration of this method.

[0162] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A low-code customer illegal funds transfer risk rating method, characterized in that: This includes receiving transaction data from customers through a unified rating interface and comparing it with pre-set risk factors to conduct illegal fund transfer risk rating; The rating configuration of the risk of illegal fund transfer is displayed and edited through a front-end visual interface, wherein the rating configuration of the risk of illegal fund transfer includes the weight and correlation strength of the risk factors.

2. A low-code customer illegal funds transfer risk rating method according to claim 1, characterized in that: The illegal funds transfer risk rating is provided with a precondition, and the precondition is provided with a precondition matching mode. If the received data satisfies the precondition matching mode, the score of the illegal funds transfer risk rating is set.

3. A low-code customer illegal funds transfer risk rating method according to claim 2, characterized in that: When the precondition matching mode is directed to a specific risk level, if the data satisfies the precondition matching mode, the final risk rating score is directly set to the preset first score.

4. A low-code customer illegal funds transfer risk rating method according to claim 2, characterized in that: When the precondition matching mode is directed to a specific risk level or above, if the data satisfies the precondition matching mode, the final risk rating score is set to at least a preset second score.

5. A low-code customer illegal funds transfer risk rating method according to claim 1, characterized in that: The illegal funds transfer risk rating is provided with a post-condition, and if the received data meets the post-condition, the final rating score or risk level is adjusted.

6. A low-code customer illegal funds transfer risk rating method according to claim 1, characterized in that: The IFT risk rating process includes: Multiple risk dimensions are preset, each risk dimension is provided with multiple risk subcategories, each risk subcategories corresponds to one of the risk factors, the risk subcategories under each risk dimension are provided with corresponding subcategories weights, and the sum of the subcategories weights under the risk dimension is 1, each risk dimension is provided with corresponding dimension weights, and the sum of all dimension weights is 1; Data matching is performed through the risk factors corresponding to each risk subcategory to obtain the score of each risk subcategory, and then the score of each risk dimension is obtained by multiplying it by the corresponding subcategory weight and accumulating it. Finally, the risk score is obtained by multiplying it by the final dimension weight and accumulating it.

7. A low-code customer illegal funds transfer risk rating method according to claim 6, characterized in that: The assignment of subclass weights to each risk subclass is based on the corresponding strength of association assignment.

8. A low-code customer illegal funds transfer risk rating method according to claim 6, characterized in that: Each risk factor is scored using a binary scoring method, a relative scoring method, or an absolute scoring method.

9. A low-code customer illegal funds transfer risk rating method according to claim 6, characterized in that: The front-end visual interface is also used to add or delete risk factors; When adding a risk factor, first determine the risk dimension to which the risk factor belongs, then select the corresponding scoring method for the risk factor, and finally assign a weight to the risk factor; When a risk factor is deleted, the subclass weight corresponding to the risk factor is set to 0.

10. A low-code customer illegal funds transfer risk rating system, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor calls the computer program to execute the steps of any one of the methods according to claims 1 to 9.

Images