A Rough Set and Evidence Theory Based Method for Fraud Detection

Due to the increase of fraud which results in economic loss of enterprises, how to diagnose fraud has become to be an active topic. In order to analyze and detect fraud, a rough set and evidence theory based method is proposed. Firstly, we employ the rough set theory to analyze the decision table which includes fraud data and abstract effective decision rules. Secondly, we regard every rule as a decision expert, and calculate their weights and belief probability assignments according to the rules and their confidences. Lastly, the evidence theory is employed to combine these belief probability assignments. The experimental results show that our method can analyze and identify fraud effectively