Predicting bad utility consumers in Malaysia

Abstract

In many organizations, especially in most utility companies including Company X, revenue collection is a major issue when customers face difficulties in paying their utility bills before the deadline. There are many reasons for this problem but it becomes a serious financial issue for the organizations when the cumulative amount of bad debts reached a staggering figure. This paper reports a study of this issue for Company X involving its customers based in Bangi and Kajang, totaling upto 1,525 customers. The study is conducted to identify the factors of customers who would default payment of their bills. The identification of such factors is important to enable Company X to identify these customers and implement the necessary measures to mitigate the problem. The CRISP-DM (Cross-Industry Standard Process for data mining) model was employed in conducting the study. The results provide an intial understanding of the issue and the solution model generated could be used to resolve the issue for other areas in Malaysia. © 2014 IEEE.