Power signal prediction by fuzzy-neural model with considering training problems

This paper introduces a new artificial neural network (NN) model, with fuzzy learning algorithm, for power signal prediction. This model is designed to take advantage of the overfitting and underfitting phenomena involved in the training of the neural networks. Results from experimental prediction data of daily power load using the proposed method and the conventional standard error back-propagation (BP) technique are presented in comparative form. Data from these preliminary experiments shows possible potential for commercial applications.