CN115689732A - Risk assessment method for anti-money laundering analysis - Google Patents
Risk assessment method for anti-money laundering analysis Download PDFInfo
- Publication number
- CN115689732A CN115689732A CN202210358231.4A CN202210358231A CN115689732A CN 115689732 A CN115689732 A CN 115689732A CN 202210358231 A CN202210358231 A CN 202210358231A CN 115689732 A CN115689732 A CN 115689732A
- Authority
- CN
- China
- Prior art keywords
- risk
- pep
- money laundering
- assessment method
- equation set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention relates to a risk assessment method for anti-money laundering analysis, which comprises the following steps: extracting PEP client-related attribute data from a plurality of data tables in which PEP client information is stored; generating a corresponding probability transfer matrix and a corresponding risk equation set according to the extracted PEP client related attribute data; solving the risk equation set to obtain a solution result; and sequencing and filtering the solving results to obtain the risk score of the PEP client. Compared with the prior art, the method comprehensively considers the relationship between the members, the relationship between the members and the influence factors and the relationship between the influence factors, can accurately calculate the risk score, and effectively identifies the PEP client with high risk.
Description
Technical Field
The invention relates to the technical field of big data processing, in particular to a risk assessment method for anti-money laundering analysis.
Background
In the financial industry, PEP customers may be involved in financial risk and money laundering behavior because they directly or indirectly obtain public funds and affect their jurisdictional administration and business transactions, and so financial institutions need to actively identify PEP customers and fully assess the specific risks that they constitute.
In the process of identifying and evaluating risks, the risk of evaluating related services matched with PEP records with higher risk is often considered preferentially according to databases compiled by different issuing organizations, and the records need to be sorted according to various category attributes and risks. The currently mainly adopted ranking algorithm is the PageRank algorithm of Google, which is the core algorithm of a Google search engine, the importance degree of a webpage is evaluated according to the mutual links among the webpages, and the ranked webpage is presented as a search result. The PageRank algorithm judges the importance degree of a node through mutual reference between the nodes in the directed network, and the importance degree of the node is higher when the number of the linked nodes is larger, the quality of the link source is higher. Problems are encountered if it is applied in the field of PEP client ranking: the relationship between records and influencing factors is not bi-directional; the relationship between records is rarely unidirectional (for example, a couple relationship), which results in low accuracy of risk assessment results and failure to identify a PEP client with high risk.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a risk assessment method for anti-money laundering analysis, which comprehensively considers the relationship between members, the relationship between members and influencing factors and the relationship between influencing factors and can accurately calculate risk scores, thereby effectively identifying the PEP client with high risk.
The purpose of the invention can be realized by the following technical scheme: a risk assessment method for anti-money laundering analysis, comprising the steps of:
s1, extracting PEP client related attribute data from a plurality of data tables in which PEP client information is stored;
s2, generating a corresponding probability transfer matrix and a corresponding risk equation set according to the extracted PEP client related attribute data;
s3, solving the risk equation set to obtain a solution result;
and S4, sequencing and filtering the solving results to obtain the risk score of the PEP client.
Further, the PEP client related attribute data comprises influence factors and mutual relations of the PEP client.
Further, the interrelationships comprise relationships among the PEP clients, relationships among the PEP clients and the influencing factors and relationships among the influencing factors.
Further, the probability transition matrix in step S2 is specifically:
0 < alpha, beta < 1 and alpha + beta =1
Wherein, alpha and beta are probability transfer coefficients, A is a relation function between PEP clients, B is a relation function between the PEP clients and influence factors, and B' is a relation function between the influence factors.
Further, the risk equation set in step S2 is specifically:
wherein, the first and the second end of the pipe are connected with each other, i risk score, y, for the ith PEP client of the m PEP clients k A risk score for the kth influencing factor of the total M influencing factors;
if political i is associated with factor k, then b ik =b′ ki =1;
further, in the step S3, a successive approximation method is specifically adopted to solve the risk equation set.
Further, the step S3 specifically includes the following steps:
s31, constructing an initial vector;
s32, multiplying the probability transition matrix by the initial vector to obtain a new current vector;
s33, comparing the current vector with the previous vector, if the current vector meets preset conditions, ending the current solving process, and outputting a solving result of the current risk equation set, otherwise, executing the step S34;
and S34, multiplying the probability transition matrix and the current vector, and returning to the step S33.
Further, each element in the initial vector is 1.
Further, in the step S33, the previous vector is subtracted from the current vector, and whether a difference between the two is smaller than or equal to a set threshold is determined, if yes, the current solving process is ended, and a solving result of the current risk equation set is output, otherwise, the step S34 is executed.
Further, the solution result of the risk equation set specifically includes:
wherein n is the current solving times.
Compared with the prior art, the method comprehensively considers the relation among the PEP clients, the relation among the PEP clients and the influence factors and the relation among the influence factors, obtains the risk equation set by generating the probability relation transfer matrix, and determines the risk score of the PEP clients by solving, sequencing and filtering the risk equation set, thereby realizing a risk evaluation scheme integrating the association of all factors, effectively identifying the PEP clients with high risk, being beneficial to investigating and auditing the transaction remittance behavior of the PEP clients and reliably reducing the money laundering risk.
According to the invention, the mutual relation among all the influence factors is introduced into risk evaluation, the probability transition matrix is generated, and a successive approximation method is adopted to solve a high-dimensional risk equation set, so that the accuracy of the solution result can be fully ensured, and further the accuracy of the risk evaluation result is ensured.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a schematic diagram of an application process in an embodiment;
fig. 3 is a schematic diagram of the relationship between PEP clients in the embodiment.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments.
Examples
As shown in fig. 1, a risk assessment method for anti-money laundering analysis includes the steps of:
the method comprises the following steps of S1, extracting PEP client related attribute data from a plurality of data tables in which PEP client information is stored, wherein the PEP client related attribute data comprise influence factors and mutual relations of PEP clients;
the mutual relations comprise the relations among the PEP clients, the relations among the PEP clients and the influence factors and the relations among the influence factors;
s2, generating a corresponding probability transition matrix and a corresponding risk equation set according to the extracted PEP client relevant attribute data, wherein the probability transition matrix specifically comprises the following steps:
0 < alpha, beta < 1 and alpha + beta =1
In the formula, alpha and beta are probability transfer coefficients, A is a relation function between PEP clients, B is a relation function between the PEP clients and influence factors, and B' is a relation function between the influence factors;
the risk equation set is specifically:
in the formula (I), the compound is shown in the specification, i risk score, y, for the ith PEP client of the m PEP clients k A risk score for the kth influencing factor of the total M influencing factors;
if political i is associated with factor k, then b ik =b′ ki =1;
s3, solving the risk equation set to obtain a solution result, wherein the risk equation set is solved by adopting a successive approximation method in the embodiment, specifically:
s31, constructing an initial vector, wherein each element in the initial vector is 1;
s32, multiplying the probability transfer matrix with the initial vector to obtain a new current vector;
s33, comparing the current vector with the previous vector (subtracting the previous vector from the current vector), if a preset condition is met (judging whether the difference value between the current vector and the previous vector is less than or equal to a set threshold value, if so, namely, the preset condition is met), ending the current solving process, and outputting a solving result of the current risk equation set:
in the formula, n is the current solving times, otherwise, the step S34 is executed;
s34, multiplying the probability transition matrix by the current vector, and returning to the step S33;
and S4, sequencing and filtering the solving results to obtain the risk score of the PEP client.
In summary, the technical scheme mainly comprises the following steps:
1. selecting related attributes from a table in which PEP client information is stored;
2. initializing the relationship and the influence factors between PEP clients into a probability transfer matrix according to a calculation formula and obtaining an equation set;
3. calculating a solution result of the equation set by using a successive approximation method;
4. and (4) sorting and filtering the results to obtain the risk score of the PEP client.
The present embodiment applies the above technical solution, wherein the process of generating the probability transition matrix and solving the risk equation set is shown in fig. 2:
1. generating a probability transfer matrix according to the influence factors and the mutual relation of the politicians;
2. generating an initial vector, wherein each element of the general vector takes 1;
3. multiplying the probability transfer matrix by the vector to obtain a new vector;
4. if the difference between the new vector and the old vector is less than or equal to a set threshold value, ending, otherwise, jumping to the step 3.
The probability transition matrix is generated in the following manner under the two influence factors:
in the formula (I), the compound is shown in the specification, i risk score, y, for the ith PEP client of the m PEP clients k A risk score for the kth influencing factor of the total M influencing factors;
if political i is associated with factor k, then b ik =b′ ki =1;
the coefficients 0 α, β < 1 and α + β =1, typically α = β =0.5.
In this embodiment, table 1 shows the relevant attribute data of the PEP client obtained from the records stored by different issuing organizations.
TABLE 1
According to the data shown in table 1, the association relationship diagram shown in fig. 3 is obtained, where the left solid line represents the interpersonal relationship, the middle dotted line represents that the issuing authority issues the information of the PEP client, and the right solid line represents that two issuing authorities issue the information of the same PEP client together.
The following set of risk equations can thus be listed:
x Zhang San =0.5*(x zhang four /2+x Wang Wu /4+y Issuing mechanism one /2+y Second issuing mechanism /5)
...
y Issuing mechanism one =0.5*(x Zhang San /4+x Li Ba /1+y Second issuing mechanism /4)
...
Finally, a ten-element linear equation set is obtained, and the unknown number is far greater than ten in practical application, so that the equation set needs to be solved by a successive approximation method:
the solution can thus be obtained:
the risk scores of PEP clients are screened from the solving result of the embodiment and then sorted, so that the highest score of Wang Wu is 1.46, the same scores of Wang Liu and Wang Xiu are parallel, the second score is 0.98, and the lowest score of Li Ba is 0.15. The Wang Wu with the highest risk is taken as a node in FIG. 3, which has the most outward connected lines, and shows that the Wang Wu is an important node in the network, and the risk score is the highest and accords with the visual feeling. According to the method, risk sequencing is carried out under the factors of the PEP client and the issuing institution, early warning evaluation can be well carried out on the risk of the politicians, and meanwhile the method can be similarly popularized to the multi-factor condition with higher dimension.
In order to perfect the risk assessment of PEP clients, screen out high-risk PEP clients, specially design a risk assessment scheme integrating the association of all factors, score and sort the records according to the mutual association among the records and the association of attributes in the records, introduce the mutual relationship among all the factors into the risk assessment field, establish information stored in a form into a series of mutual association relationship functions, generate a probability transfer matrix and a high-dimensional risk equation set, and combine the solving mode of risk scores under the high-dimensional condition, thereby fully ensuring the accuracy of the risk assessment.
Claims (10)
1. A risk assessment method for anti-money laundering analysis, comprising the steps of:
s1, extracting PEP client related attribute data from a plurality of data tables in which PEP client information is stored;
s2, generating a corresponding probability transfer matrix and a corresponding risk equation set according to the extracted PEP client related attribute data;
s3, solving the risk equation set to obtain a solution result;
and S4, sequencing and filtering the solving results to obtain the risk score of the PEP client.
2. The risk assessment method for anti-money laundering analysis according to claim 1, wherein said PEP client related attribute data comprises influence factors and interrelationships of PEP clients.
3. The risk assessment method for anti-money laundering analysis according to claim 2, wherein said interrelationships comprise relationships between PEP clients, relationships between PEP clients and influencing factors, relationships between influencing factors.
4. The risk assessment method for anti-money laundering analysis according to claim 3, wherein the probability transition matrix in step S2 is specifically:
0 < alpha, beta < 1 and alpha + beta =1
Wherein, alpha and beta are probability transfer coefficients, A is a relation function between PEP clients, B is a relation function between the PEP clients and influence factors, and B' is a relation function between the influence factors.
5. The risk assessment method for anti-money laundering analysis according to claim 4, wherein the risk equation set in step S2 is specifically:
wherein x is i Risk score, y, for the ith PEP client of the m PEP clients k A risk score for the kth influencing factor of the total M influencing factors;
if political i is associated with factor k, then b ik =b′ ki =1;
6. the risk assessment method for anti-money laundering analysis according to claim 5, wherein said step S3 is implemented by solving the risk equation set by using successive approximation.
7. The risk assessment method for anti-money laundering analysis according to claim 6, wherein the step S3 comprises the following steps:
s31, constructing an initial vector;
s32, multiplying the probability transfer matrix with the initial vector to obtain a new current vector;
s33, comparing the current vector with the previous vector, if the current vector meets preset conditions, ending the current solving process, and outputting a solving result of the current risk equation set, otherwise, executing the step S34;
and S34, multiplying the probability transition matrix and the current vector, and returning to the step S33.
8. The risk assessment method for anti-money laundering analysis according to claim 7, wherein each element in said initial vector is 1.
9. The method as claimed in claim 7, wherein the step S33 is to subtract the previous vector from the current vector, determine whether the difference between the two is less than or equal to the predetermined threshold, if so, terminate the current solution process, output the solution result of the current risk equation set, otherwise, execute step S34.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210358231.4A CN115689732A (en) | 2022-04-06 | 2022-04-06 | Risk assessment method for anti-money laundering analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210358231.4A CN115689732A (en) | 2022-04-06 | 2022-04-06 | Risk assessment method for anti-money laundering analysis |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115689732A true CN115689732A (en) | 2023-02-03 |
Family
ID=85060493
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210358231.4A Pending CN115689732A (en) | 2022-04-06 | 2022-04-06 | Risk assessment method for anti-money laundering analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115689732A (en) |
-
2022
- 2022-04-06 CN CN202210358231.4A patent/CN115689732A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2019196549A1 (en) | Method and device for determining high-risk user | |
CN111291816A (en) | Method and device for carrying out feature processing aiming at user classification model | |
US20210182859A1 (en) | System And Method For Modifying An Existing Anti-Money Laundering Rule By Reducing False Alerts | |
US7818322B2 (en) | Efficient method for clustering nodes | |
US20120150825A1 (en) | Cleansing a Database System to Improve Data Quality | |
US10049155B2 (en) | System for mending through automated processes | |
Tan et al. | Time series clustering: A superior alternative for market basket analysis | |
CN111709826A (en) | Target information determination method and device | |
US20210271666A1 (en) | Analyzing a processing engine of a transaction-processing system | |
Brodeur et al. | Unpacking p-hacking and publication bias | |
WO2016095727A1 (en) | Method and device for establishing category mappings | |
CN115545886A (en) | Overdue risk identification method, overdue risk identification device, overdue risk identification equipment and storage medium | |
CN111833182B (en) | Method and device for identifying risk object | |
CN115689732A (en) | Risk assessment method for anti-money laundering analysis | |
CN113870021B (en) | Data analysis method and device, storage medium and electronic equipment | |
US20240152818A1 (en) | Methods for mitigation of algorithmic bias discrimination, proxy discrimination and disparate impact | |
CN112823502B (en) | Real-time feedback service for resource access rule configuration | |
US20210073247A1 (en) | System and method for machine learning architecture for interdependence detection | |
US20160203220A1 (en) | Method and apparatus for natural language searching based on mccs | |
US10636083B1 (en) | Systems methods and articles of manufacture for analyzing on-line banking account data using hybrid edit distance | |
CN115082079B (en) | Method and device for identifying associated user, computer equipment and storage medium | |
CN116167852B (en) | Method and device for processing fund flow direction data | |
Tafvizi et al. | Attributing AUC-ROC to Analyze Binary Classifier Performance | |
CN110109994A (en) | Auto metal halide lamp air control model comprising structuring and unstructured data | |
Brito et al. | A framework to improve churn prediction performance in retail banking |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |