Fraud Detection Using a Fuzzy Expert System In Motor Insurance

Abstract

Insurance industry experts believe that fraud is a destructive disaster in the insurance industry. Over the years, many methods have been used in the literature for fraud detection, one of which is expert systems. Fraud detection expert systems are based on the knowledge of experts in the field of insurance identify fraud. Judgment of experts is mostly based on evidence, documents, qualitative information which is often presented in verbal words to describe the fraudulent behavior. In the presented model, 61 qualitative and quantitative criteria related to the detection of fraud in car insurance were identified. Then, these criteria were prioritized according to expert opinion and 17 criteria with the highest priority classified into eight factors were selected. In the suggested system fuzzy inference was performed using Mamdani algorithm. Finally, the designed system was implemented to an Iranian private insurance company and the validity of the system assessed by a questionnaire and came up to 69.45%. The obtained results indicate that the proposed model is able to detect the fraud quite significantly.