Using Elman and FIR neural networks for short term electric load forecasting

Abstract

Finite impulse response (FIR) neural network and Elman neural network have been compared in electric load prediction. An FIR neural network has been trained with a temporal back-propagation learning algorithm and the results obtained showed that the effectiveness of the algorithm is more important than the applied network model. The comparison between both networks and the standard approach with Multilayer perceptron (MLP) network, demonstrates that the FIR network acts adequately. It performs better than the Elman network. Both networks perform better than the MLP network.