Data Mining using K-Means Cluster Method to group policyholder of motoring vehicles in Indonesia.

Abstract

Data mining is a process to determine interesting pattern from bigger data. One technique of data mining is K-means cluster. This method used to group the data collection into a K cluster so data points in a cluster is similar on one to each other rather than data points of different cluster. This study decribes K-means cluster method used to group policyholders  of motoring vehicles assurance in Indonesia based on variables of insurance code, vehicle code, use code, zone code, vehicle age, and insurance price. When the cluster number is five, so cluster 1 has members about 826 sed on policyholders, cluster 2 has members about136 policyholders, cluster 3 has 522 policyholderss, cluster 4 has 442 policyholders, cluster 5 has 2.151 policyholders. Cluster 1 upto cluster 5 are dominated by policyholders who take assurance product with comprehensive assurance, the policyholders who use Toyota brand, policyholders who use owncars, policyholders whose vehicles are in Jabodetabek area..