Gath-geva approach to forecast electric energy consumption

Short-term load forecasting is necessary for adequate scheduling and operation of power systems. It's generally made by developing models in relations to climate and previous load data. In this paper, we discuss in detail how Fuzzy clustering based on Gath and Geva algorithm is successfully applied to electric load forecasting. Results and discussions from real-world case studies based on data from RTE France of electricity consumption in 2010 are presented. The variance accounted for (VAF) and the RMSE indices used in modeling process are respectively 99.5059 and 0.0545 for training and 85.8950, 0.1872 for validation showing the good prediction performance and the suitability of the proposed approach for short-term electric load forecasting.