HYBRED BETWEEN RADIAL BASIS FUNCTION NETWORK AND ARIMA MODELS FOR FORECASTING ELECTRICITY CONSUMPTION IN ALGERIA FOR 1980-2019

Abstract

This research aims to use hybrid models that combine ARIMA models in the estimation of linear variations and radial base function network models for the estimation of non-linear variations, In order to estimate it, to predict it and to compare it with ARIMA models, In order to compare the models used, MSE and RMSE were applied. The time series data of electricity consumption in Algeria were used for the period (1980-2019), and it was found that hybrid models are better than ARIMA models for forecasting electricity consumption in Algeria.