Identifying the effective factors in the profit and loss of vehicle third party insurance for insurance companies via data mining classification algorithms

Abstract

Background: Insurance companies for surviving and keeping the market, always emphasize on profitability and reduction of their losses. Methods: Investigation of the Car Insurance Information shows that main factors which effect on profit or loss of insurance companies include: type of vehicle usage, license, type of license and compliance or lack of compliance with the vehicle, the amount of the premium, rate of the commitment, the car quality of the car companies, the age of the driver, driver education, the mismatch premiums with the insurance case, the delay in the renewal of insurance policies. Findings: In this paper, using data mining information of 2011year of third party insurance in Iran insurance companies in Kohgiluyeh and Boyer Ahmad province are studied. The results showed that using the classification algorithm with over 91% accuracy and decision trees with over 96% accuracy able to provide a model to identify effecting factors and determine their impact on the profit and loss of vehicle third party insurance. Applications/Improvements: Comparing results indicated that, the decision tree Wj48 with very high accuracy will able to detect and predict the occurrence of the damage of an insurance, properly. After that two other algorithms as well as with high accuracy and the same have this ability.