isk Based Comparison between two data mining methods in segmentation of car insurance customers (Case Study: Mellat Insurance Company

Abstract

Due to the sharp rise of the information technology (IT), the amount of datastored in databases is dramatically on the rise. Analyzing the stored data andconverting it to information and knowledge which is applicable in organizationsrequires powerful instruments. As with other economic sectors, recognizing andattracting low-risk and profitable customers are of high significance for insuranceindustry. Car insurance is one of the most important insurance brancheswhich accounts for a great deal of portfolio of insurance industry. Risk segmentationof policyholders on the basis of observable features can help insurancecompanies to reduce loss, raise the rate of insurance coverage, and prevent themfrom making an inappropriate choice in the insurance market. In this study, thesegmentation of comprehensive car insurance customers on the basis of risk wasselected through self-organizing map and K-means. At first, the effective factorson the risk of policyholders are identified. Then, the insurance policyholders aresegmented using the proposed SOM and K-means. Customers’ characteristicsin every cluster are identified. Finally, the two methods compared with eachother. The advantages and disadvantages of them illustrated

Keywords

• customer segmentation
• self-organizing map
• k-means
• comprehensive auto insurance