Medium-Term Electric Energy Demand Forecasting Using Generalized Regression Neural Network

Abstract

Medium-term electric energy demand forecasting is becoming an essential tool for energy management, maintenance scheduling, power system planning and operation. In this work we propose Generalized Regression Neural Network as a model for monthly electricity demand forecasting. This is a memory-based, fast learned and easy tuned type of neural network which is able to generate forecasts for many subsequent time-points in the same time. Time series preprocessing applied in this study filters out a trend and unifies input and output variables. Output variables are encoded using coding variables describing the process. The coding variables are determined on historical data or predicted. In application examples the proposed model is applied to forecasting monthly energy demand for four European countries. The model performance is compared to performance of alternative models such as ARIMA, exponential smoothing, Nadaraya-Watson regression and neuro-fuzzy system. The results show high accuracy of the model and its competitiveness to other forecasting models.