DBSCAN Clustering Algorithm Applied to Identify Suspicious Financial Transactions

Money laundering refers to disguise or conceal the source and nature of variety ill-gotten gains, to make it legalization. In this paper, we design and implement the anti-money laundering regulatory application system (AMLRAS), which can not only automate sorting and counting the money laundering cases in comprehension and details, but also collect, analyses and count the large cash transactions. We also adopt data mining techniques DBSCAN clustering algorithm to identify suspicious financial transactions, while using link analysis (LA) to mark the suspicious level. The presumptive approach is tested on large cash transaction data which is provided by a bank where AMLRAS has already been applied. The result proves that this method is automatable to detect suspicious financial transaction cases from mass financial data, which is helpful to prevent money laundering from occurring.