Research On Data Mining-Based Suspicious Money Laundering Transactional Behavioral Patterns Recognition

The purpose of money-laundering is to realize the legalization of crime receipts. Since the money-laundering offenses can support other crimes, it is named not only as a "Lifeline" for its downstream crime, but also as a "umbrella" for the criminal liability of its upstream crime. Money-laundering activities will disrupt normal economic and financial order as well as disrupt social stability. Especially in China, money-laundering offenses will fuel the spread of corruption. Therefore, all countries in the world take active measures to combat money-laundering offenses. The promulgation and implementation of "Anti-Money Laundering" in China indicates a new stage of anti-money laundering.At present, China's anti-money laundering work mainly regards the suspicious transactions data, which are reported based on the regulations in the《Management Measures》for grassroots financial institutions, as a base to research money-laundering offenses. But this suspicious transaction reporting system has some problems, such as fuzzy criterions, massive data, a high false alarm rate, lack of adaptability and easy to avoid, which will affect the reliability and validity of the reported suspicious transactions data. The collection and analysis of the data of irregular transaction patterns has been recognized as the core issue against money-laundering crimes. Therefore, this paper puts forward that the account historical transaction data were analyzed to identify abnormal transactions pattern in the accounts through constructing suspicious money laundering transactional behavioral pattern recognition method. These unusual transactions pattern data can be regarded as the suspicious transactions data to report.Data mining technology can rapidly handle a great deal of financial data in order to identify money laundering patterns, which will make anti-money laundering process more simple and efficient Therefore, this paper takes the real transaction data of financial institutions' accounts and the artificial simulation data, which is formed according to the real data, as the research object. Using data mining techniques constructs a time , labor saving and effective suspicious money-laundering transactional behavioral pattern recognition methods. That will provide more reliable data for the Anti-Money Laundering Monitoring and Analysis Center of the People's Bank and help to increase the quality and efficiency of anti-money laundering monitoring and investigationThis paper firstly analyzes the development process of the world's anti-money laundering system. Based on the research literatures from both home and abroad, it points that the reporting system for identifying suspicious transactions in reported transaction data is a base of China's anti-money laundering work, which can determine the significant and purpose of this study and research. Secondly, by analyzing China's anti money-laundering work and money-laundering theory in detail, mainly by the problems in the anti-money-laundering work, this article indicates the necessity of identifying suspicious money-laundering transactional behavioral pattern. Thirdly, after comparative analysis for various data mining technology in the identification of suspicious money-laundering transactional behavioral pattern, this paper points out the best clustering algorithm and outlier detection algorithm for suspicious behavioral pattern recognition pattern of money-laundering transactions. Based on these two algorithms, this article constructs a improved CBLOF algorithm, on which it builds the suspicious money-laundering transactional behavioral pattern recognition algorithm. Finally, adopting C + + programming, this paper analyzes the real transactions data and the artificial simulation dataThe main contributions of this study are: (1) Using improved CBLOF algorithm selects the four variables (transaction amount, dispersion coefficient of the transaction amount and accounts withdrawal frequency, and deposit frequency) to order suspicious degree of the historical accounts' money laundering transactions, which will provide high-quality suspicious transactions data to the grassroots financial institutions; (2) Using artificial simulation data generation algorithm makes transaction data containing laundering information to settle, on a certain extent, the problem of lackingAnti money-laundering case study data.