Discovering Hidden Group in Financial Transaction Network Using Hidden Markov Model and Genetic Algorithm

Financial crimes such as money laundering are often committed by cooperative individuals in a hidden manner. Discovering hidden group in financial transaction networks can help to find suspects of money laundering. A method is presented to discover the hidden group based on hidden Markov model (HMM) and genetic algorithm. HMM is used to describe financial transaction network. The maximum likelihood principle is adopted to transform hidden group detection to a combinational optimization problem. An effective genetic algorithm is devised to solve the optimization problem according to the characteristic of the feasible solutions. Real financial transaction data is preprocessed by considering multi-relations among the accounts. Effectiveness and efficiency of our method is validated by experiments on both synthetic and real dataset.