Neural Networks as Radial-Interval Systems through Learning Function

Abstract

This paper presents a novel dimension of neural networks through the approach of interval systems for great more forecasting activity. The artificial neural network (ANN) based models are the most popular ones for load forecasting and other applications. This approach is only a thought line that can enhance the fundamental requirement of all networks giving and incorporating the analytical and expertise knowledge in forecasting from the existential approaches. [1] Interval systems as an approach to approximate interval models by neural networks is proposed.