Forecasting Electricity Consumption with Neural Networks and Support Vector Regression

Abstract

Energy strategy is extremely important for developing countries. As the economy of these countries grow rapidly, their energy consumptions increase substantially. Turkey's high growth rate in the last decade resulted with significant increase in energy consumption. Policy makers should give critical decisions and develop new strategies for meeting this growing energy demand. Apparently, accurate predictions of the future energy consumption are vital for developing such strategies. In this study, we have used the state of the art computation methods to forecast the electricity consumption of Turkey. The forecast results are compared with real consumption values to measure the performance of the methods.