Performance Evaluation of Naïve-Bayes Classifier in Predicting the Energy Not Served (ENS) Class of Electric Power Distribution System

Customer satisfaction is an important factor in the electrical system business. In order to improve customer satisfaction, the Energy Not Served (ENS) in the electric power distribution system which affects to the poor reliability index must be reduced maximally. Among the ENS reduction strategies, Machine Learning (ML) also plays some main roles such as in predicting the ENS class due to certain fault in the electric power distribution system. This paper is presenting the performance measurement of the Naive-Bayes Classifier (NBC) in predicting the ENS class of electric power distribution system by adding gain ratio feature selection technique which produces a fair performance with accuracy and average specificity value about 77-84%.