Predicting Fraudulent Motor Claims for Insurance Industry

Abstract

"Overall focus of this project is to analyze the past records and identify the common pattern which can be used to predict the motor claims reported by the customer, whether it is genuine or not. When there is a lot of fraud claims have been requested in an insurance company and if it is not identified correctly, it causes an adverse impact financially and it may cause damage to the reputation of the company. The goal of this research is to develop a model which can help the insurance company to identify the genuine motor claims using computerized claims related details rather than following the manual method. During the study we have collected real-time data of motor claims of an insurance company and conducted a detailed analysis on the data to finalize the dependent and independent variables. Based on the variables, different algorithm is used to identify the most accurate model to be used for this problem. Based on the outcomes of different mode, Logistic Regression is applied for this project to get the high prediction accuracy. This research helps in identifying factors which impacts the process of predicting the genuine claims and fraud claims. This model promises a reliable outcome which helps in the day-to-day business process of the insurance claiming process. This model helps in reducing the time taken to process the claims. Currently SMEs are required to analyze the genuineness of the claim manually which will be take more time since in a day they will receive hundreds of claim request based on the size of the insurance company. Through this model SME is not expected to preform it, the whole process will be automated therefore time consumption is less and resource can be used for the business growth than performing this monotonous work."