Improving the performance of artificial immune system in estimation problems with normalization technique: A case study of USA, Japan and France electricity consumption

Abstract

This paper presents an artificial immune system (AIS) for electricity consumption estimation as a common problem in estimation domain. We study the impact of data normalization on artificial immune system (AIS) performance and two hundred AIS are constructed for this. Also, fifty AIS have been constructed and tested in order to finding best AIS for electricity consumption estimation in each case. Another unique feature of this study is the utilization of AIS in estimation domain and especially in electricity consumption estimation as the first time. Two standard inputs are used in order to training and testing developed AIS. The mentioned input parameters are gross domestic product (GDP) and population (POP). All of trained AIS are then compared with respect to mean absolute percentage error (MAPE). To meet the best performance of the intelligent based approaches, data are normalized. To show the applicability and superiority of the AIS, actual electricity consumption in USA, Japan and France from 1980 to 2007 is considered.