Abstract :
Banks are faced with the challenge of adapting to the ever-changing risk and methods related to money laundering. Money laundering is a complex, dynamic and distributed process. Some Anti-Money Laundering (AML) Systems simply transforms vast quantities of data into vast numbers of reports that do not facilitate timely detection or effective interdiction. Transaction Flow Analysis (TFA) system is proposed to get over this issue. The main parts of this TFA system is, first, bank transaction importer, which is not bound to any file format. Second, application of frequent pattern mining and transaction mining algorithms to detect money laundering, which implements distributive box and collective box. Third, clustering the transaction and detecting suspicious clusters by recognizing laundering methods and roles of the offender. Finally obtained clusters and frequent patterns can be visualized in schema and timeline diagrams. It allows the bank or police analyst to find suspected money flows and suspicious transactions.
Keywords :
bank data processing; data mining; pattern clustering; risk management; security of data; transaction processing; AML systems; TFA system; anti money laundering systems; complex process; distributed process; dynamic process; frequent pattern mining algorithm; money laundering detection; schema; suspicious cluster detection; timeline diagrams; transaction flow analysis system; transaction mining algorithm; Transaction analysis; clustering; frequent pattern mining; suspicious transaction; visualization;
