Short-term load forecasting using BiLinear recurrent neural network

Abstract

A prediction scheme of short-term electric load forecasting using a BiLinear Recurrent Neural Network (BLRNN) is proposed in this paper. Since the BLRNN is based on the bilinear polynomial, it has been successfully used in modeling highly nonlinear systems with time-series characteristics and the BLRNN can be a natural choice in predicting electric load. The performance of the proposed BLRNN-based predictor is evaluated and compared with the conventional MultiLayer Perceptron Type Neural Network (MLPNN)-based predictor. Experiments are conducted on load data from the North-American Electric Utility (NAEU). The results show that the proposed BLRNN-based predictor outperforms the MLPNN-based one in terms of the Mean Absolute Percentage Error (MAPE). © Springer-Verlag Berlin Heidelberg 2007.