A short-term temperature forecaster based on a novel Radial Basis Function Neural Network

摘要

Many applications of building electric load forecasting require short-term hourly temperature prediction. A new method of short-term temperature forecasting based on a Radial Basis Function Neural Network, which is initialized by a Regression Tree, is presented. In this method, each terminal node of the tree contributes one hidden unit to the RBF network. Thus, the tree sets the number, position and sizes of all RBFs in the network. From a RBFNN structure point of view, the number of centres, set in that form, is very small; so it is faster to compute the weights. The forecaster uses the current coded hour and the temperature as inputs, and predicts the next hour temperature. A short database updated daily, is used for training the network. The results of this method are encouraging, therefore, it can be used as a tool for load forecasting.