A DATA MINING APPROACH FOR DETERMINING POWER CONSUMPTION OF ETHIOPIAN ELECTRIC UTILITY CUSTOMERS

Abstract

Electric industry is one of most important service provider and back bone of the energy sector in the world. Ethiopia electric utility is the only national organization distributing electric power in our country. Electric power industries are being pushed to and quickly respond to the individual and organization needs and wants of their customers due to the dynamic and highly competitive nature of the industry. According to Energy pedia published in 2016, only 27 % of the population in Ethiopia has access to electricity grid The aim of this study is designing a predictive model for determining power consumption of Ethiopian electric utility customers using data mining techniques. This study conducted in Ethiopian electric utility customers to mining big data. The approach followed in this research is hybrid data mining methodology, which being able to be the classification of customer based on power consumption, and to develop a prediction model using classification algorithms. The major steps followed are problem understanding, data understanding, Data Preparation, Modeling, evaluation of knowledge discovering and design user interface to use the discovered knowledge. The data covers from January 2008 to January 2011 E.C for all Ethiopian utility customers data included. The data prepared for mining contain 14 attribute with 85,849 instances. The study has used four classification algorithms to build predictive model namely: J48, bagging, random tree and PART. The result obtained from the experiments showed that J48 algorithm performed best with accuracy of 96.61% than the other models. In this model the number of correctly classified instances is 82,939 (96.61%) and the number of incorrectly classified instances is 2,910 (3.38%). This study has been classification of prediction power consumption based on new connection of electric utility customers either high and low power consumption. Hence, based on the findings of this study, the researcher would like to forward recommendations for electric industry to conduct the study further and come up with system that enable to an optimal management of power consumption.