Anti-money Laundering (AML) Research - A System for Identification and Multi-classification.
Yixuan Feng,Chao Li,Yun Wang,Jian Wang,Guigang Zhang,Chunxiao Xing,Zhenxing Li,Zengshen Lian
DOI: https://doi.org/10.1007/978-3-030-30952-7_19
2019-01-01
Abstract:Anti-money laundering (AML) is of great significance for the integrity of financial system in modern society because of the huge amounts of money involved in and close relationships with other types of crimes. The typical AML system includes FinCEN for America which began in 1993 and AUSTRAC for Australia. This paper aims to develop a suspicious behavior detection and categorization system based on money transaction data collected from the simulator called Paysim. For preparation work, in order to facilitate crimes classification in the second step, the associated criminal activities were divided into five categories according to their different characteristics shown in the money transaction process. Then on the basis of the transaction data, a user profile was created and new features concerning both individual parties and network effect were extracted from the profile. With combined features, two models were developed for detection and classification using supervised learning methods separately. The results show good accuracy and recall rate which are most valued in reality. Meanwhile, the models display good robustness to further adjustment for practical use. Finally two models were connected in series, and the result shows a relatively good overall performance and verifies the feasibility of the system, as it provides more choices for users to decide which model (or both) to apply according to different situations in practice.