Using Data mining for GEDCO customers' data to increase the collection of electricity bills in Gaza Strip

Abstract

The Gaza Governorate Electricity Distribution Company (GEDCO) is an economic service company operating with a special financial liability and owned by the Palestinian National Authority and local authorities.The unpaid bills are one of the company's biggest operating obstacles that obstructs money processing. The ever-increasing number of unpaid invoices has caused constant problems with the company's cash flow.In this paper, Data mining techniques and neural networks techniques were implemented over the data warehouses of GEDCO. Decision tree machine learning algorithm is used to classify the customers according to their commitment to payment, as well as to extract the categories for each class of customers. The neural networks are used to train the machine to extract a threshold for consumption and amperes. The threshold extracted for each subscriber according to his history of consumption and payment, to help him stay committed to payment.The results concluded that consumption should not exceed some value, based on the address of the subscribers, the number of amperes and the history of payment.The study that will be recommended to GEDCO, should improve their data warehouses to include more features and categories of their databases especially the subscribers’ occupations and level of income.