Examination of Vehicle Fraud Detection Possibilities with the Help of Fuzzy Inference System

Abstract

Insurance fraud is when a person or legal entity seeks to gain an improper advantage by making an incorrect compensation claim. These cases can cause severe economic damage. As a result, the detection of fraudulent incidents is an important issue nowadays, specifically in the case of the liability motor insurance market. In the past decades, soft computing techniques have emerged to model and support the recognition of the problem. In this paper, a theoretical Mamdani-type Fuzzy inference system is introduced to predict the assumed probability of being an insurance fraud with the help of easily determinable parameters: the insurance payout, Ft; the age of innocent participant vehicle, years; and the payment period of the insurance contract. The output variable of the system generated was the assumed probability, %. The model aims to describe critical events and detect suspicious cases at an early stage.