Graph mining approach to suspicious transaction detection

摘要: Suspicious transaction detection is used to report banking transactions that may be connected with criminal activities. Obviously, perpetrators of acts strive make the as innocent-looking possible. Because activities such money laundering involve complex organizational schemes, machine learning techniques based on individual analysis perform poorly when applied suspicious detection. In this paper, we propose a new method for mining graphs. The proposed in paper builds model subgraphs contain transactions. our parametrized using fuzzy numbers which represent parameters and detected. transferring through variable number accounts representing also respect some structural features. contrast other graph methods isomorphisms are match data model, matching structures.