Chaotic time series prediction with multi-layer perception and meta-heuristic methods

Abstract

Predicting chaotic time series is an applicable issue so many scientists have introduced different methods to predict their behavior. Artificial neural networks are a tool that forecasts system behavior. These tools should be trained and backpropagation algorithms used as learners. But this training process may be falling into trap of local optimum. Heuristic methods are introduced to solve this challenge.