PLNのプリペイド電気利用顧客のファジーc-平均法を用いたクラスタリングについて

抄録

Electricity Prepaid is a new customer service of Indonesian national electric company PLN. In CRM, it is said that analyzing the customer’s behavior is the most efficient way to enhance their satisfaction. In this research, we apply Fuzzy c-Means (FCM) algorithm to the house hold customer data of PLN to cluster the customer behavior. We employ XB index to validate the clustering result. As a result, we conclude that attributes of household customer are more accurate if they are clustered into 4 clusters, while in number of family member attribute is more accurate if clustered into 5 clusters.