Evolutionary graph mining in suspicious transaction detection
Evolutionary graph mining in suspicious transaction detection

Summary/Abstract: Money laundering may involve complex organizational schemes designed to obfuscate the real purpose of money transfers. In this paper, we present a graph mining method that allows detection of transaction subgraphs containing suspicious transactions. Suspicious subgraph model is parameterized using fuzzy numbers which represent parameters of transactions and some structural features of the transaction subgraphs itself. The method presented in this paper uses fuzzy matching of graph structures which allows detecting money-laundering schemes which differ to some extent from those annotated by an expert.

• Details
• Contents