A Fraud Gang Identification Method Based on Modularity and Balanced Label Propagation

Abstract

The invention discloses a fraud gang identification method based on modularity and balanced label propagation, which includes: using ID features combined with the fraud identification known by the user itself, calculating pairwise similarity for all users, establishing a similarity matrix, and using the similarity Create an association graph from the matrix; run the Louvain algorithm on the established graph to obtain the community and level information to which each node belongs; use the community, level information, and fraud identification of each node as the initial community information of each node, and run balanced label propagation The process obtains the final community to which each node belongs, and then divides the network according to whether it belongs to the common community, and divides the fraudulent gangs according to the fraud identification obtained through propagation. For the first time, the present invention applies the fraud gang identification method based on modularity and balanced label propagation to the fields of application anti-fraud and transaction anti-fraud, uses information such as transaction association to construct a correlation graph, integrates community modularity information, and uses balanced label propagation algorithm to detect fraudulent communities , to guard against potentially fraudulent transactions.

Description

technical field [0001] The invention belongs to the fields of transaction anti-fraud and application anti-fraud, and in particular relates to a fraud group identification method based on modularity and balanced label propagation Background technique [0002] With the explosive growth of online businesses such as e-commerce and third-party payment, online fraud cases are becoming more and more rampant, and showing a trend of changing techniques and fields. How to effectively and timely identify frequent online frauds Behavior has become a pressing issue. Traditional online fraud detection methods usually model each online transaction or merchant entity, and implement the construction of relevant characteristics of business flow for fraud detection. This method has an excellent effect on fraudulent behavior with obvious characteristics of the transaction itself, but it ignores the The gang association behind the fraudulent transaction is poor in identifying the fraud of the g...

AI Technical Summary

Problems solved by technology

Traditional online fraud detection methods usually model each online transaction or merchant entity, and implement the construction of relevant characteristics of business flow for fraud detection. This method has an excellent effect on fraudulent behavior with obvious characteristics of the transaction itself, but it ignores the The gang connection behind the fraudulent transaction, and the fraud detection ability of the gang that forges normal user information is poor

Images