Money Laundering Detection And System Implementation Based On Density Measurement

Money laundering is the illegal process of concealing the origins of money obtained illegally by passing it through a complex sequence of banking transfers or commercial transactions.Money laundering is often related to criminal behavior such as drug trading and terrorism,which seriously affects the harmony and stability of society.So detecting money laundering has great socioeconomic significance.The criminals often launder money utilizing interbank transactions,avoid being tracking or analyzing by the bank.Existing dense subgraph detection algorithms often focus on detecting dense subgraph on bipartite graph,which can only detect abnormal behaviors existing during the direct transfers.However,interbank money laundering involves transfers between multiple accounts,so how to detect interbank money laundering is a urgent problem that needs to be solved.In the process of money laundering,a large amount of funds pass through several layers of specific bank accounts,forming a high-density flow subgraph,which is significantly different from the density subgraph of bipartite graph formed by normal transactions.Therefore,this paper focus on studying the maximum flow density subgraph to detect inter-bank money laundering behavior.Specifically,this paper proposes a maximum flow subgraph detection algorithm,FlowScope,based on a multipartite graph model.It detects the abnormal density flows in the multi-partite graphs to mine cross-bank money laundering behaviors.First,a multipart graph model is generated from the bank's transaction data to describe the abnormal behavior of funds transferred to the intermediate bank,and then transferred from the intermediate bank to the other banks.By analyzing the characteristics of funds flow in the inter-bank money laundering process,this paper designed a abnormality measure in the form of a maximum flow density subgraph to distinguish direct normal transactions and chain money laundering behaviors.Finally,the FlowScope algorithm is used to find the largest flow density subgraph by continuously deleting the nodes that contribute little to the abnormal metric value increasing the abnormal metric value of the remaining subgraphs.Therefore,the FlowScope algorithm can capture the complete dense flow of the money laundering process,from the source accounts to the middle accounts to the destination accounts by detecting the maximum flow density subgraph,avoiding the false alarm of the bipartite density subgraph formed by direct normal transactions,which improves the accuracy of money laundering detecting.A variety of comparative experiments were performed on two real bank transaction data sets,and the experimental results show that FlowScope is effective in detecting inter-bank money laundering behavior.The main innovations of this paper are:1.Describe chained transactions using multipartite graphs:This paper uses the multipartite graph model to describe the complete process of fund during chain transfer transactions,from source accounts to intermediate accounts to destination accounts,providing novel tools for detecting inter-bank money laundering.2.novel money laundering metric:This paper analyzes the characteristics of money flow in the process of money laundering,and summarizes two anomalous characteristics in money laundering behaviors,then this paper designs an anomaly metric for inter-bank money laundering behavior in the form of the maximum flow density subgraph.which distinguishes inter-bank money laundering and normal transactions effectively.Then,in order to solve the problem of detecting inter-bank money laundering behaviors,this paper analyzes the requirements of banks' anti-money laundering scenarios,and develops an anti-money laundering system through system architecture design,system module design and database system design step.This system integrates a variety of anomaly detection algorithms,and through visual display of the detected abnormal accounts,it provides better support for anti-money laundering staff to determine the suspiciousness of these abnormal accounts.This anti-money laundering system effectively detects and analyzes abnormal transaction behaviors in banks,and fully meets the business needs of bank's anti-money laundering.