Detecting Fraud Patterns in Telecommunications Using Case Based Reasoning

All round the world, fraud situations are significantly causing huge revenue leakage in the telecommunication companies every year. The reuse of previous cases is an important issue in dealing with fraud pattern in a data with string features. Case Based Reasoning (CBR) systems have a set of cases inform of library used to facilitate the process of validation of new cases without the direct involvement of a domain expert. The proposed detection technique in this paper is based on Case Based Reasoning used to detect the occurrence of fraud with a meaningful confidence in telecommunication data. Experimental result on the fraud data indicates that the weight for all attribute used in this study needs to be set as 0.1 in order to get 98% similarity performance.