Application of Preconditioned RBFN to Temperature Forecasting for Short-term Load Forecasting

This paper proposed an efficient method for temperature forecasting for short-term load forecasting in power systems. It is well-known that as an input variable, temperature is one of the most important variables that affect a short-term load forecasting model significantly. In practice, it is important to forecast temperature precisely in dealing with short-term load forecasting. In this paper, a preconditioned ANN-based method is proposed to improve the model accuracy of temperature forecasting. As a precondition technique, deterministic annealing (DA) is used to classify input data into some clusters. The radial basis function network (RBFN) is employed as ANN at each cluster so that one-step ahead temperature is evaluated precisely. The effectiveness of the proposed model is demonstrated for real data