A clique-based method for mining fuzzy graph patterns in anti-money laundering systems
Л. С. Берштейн, Alexey Tselykh · 2013
In this paper, we present a clique-based method for mining fuzzy graph patterns of money laundering and financing terrorism. The method will contribute to a new generation of intelligent anti-money laundering systems that incorporate comprehensive information from various information sources as well as from human subject matter experts. A fuzzy degree of confidence can therefore be associated to each relation between any two actors in a graph of transactions. We reduce the problem of fuzzy subgraph isomorphism to the problem of finding a fuzzy set of maximal cliques in a fuzzy extension of a mathematical construct that is similar to Vizing's modular product of graphs.