Integration of multi layer perceptron and design of experiments for forecasting household electricity consumption

Abstract

Due to various seasonal and monthly changes in electricity consumption, it is difficult to model it with conventional methods. This paper illustrates an Artificial Neural Network (ANN) approach based on supervised multi layer perceptron (MLP) network for household electricity consumption forecasting. This is the first study which uses MLP for forecasting household electricity consumption. Previous studies base their verification by the difference in error estimation. However this study shows the advantage of MLP methodology through design of experiment (DOE). Moreover, DOE is based on analysis of variance (ANOVA) and Duncan Multiple Range Test (DMLT). Furthermore, actual data is compared with ANN-MLP and conventional regression model. The significance of this study is integration of MLP and DOE for improved processing, development and testing of household electricity consumption. Moreover, it would provide more reliable and precise forecasting for policy makers. To show the applicability and superiority of the integrated approach, annual household electricity consumption in Iran from 1974 to 2003 was collected for processing, training and testing purpose.