Forecasting using non-linear techniques in time series analysis : an overview of techniques and main issues

Abstract

The development of techniques in non linear time series analysis has emerged from its time series background and developed over the last few decades into a range of techniques which aim to fill a gap in the ability to model and forecast certain types of data sets such a chaotic determinate systems. These systems are found in many diverse areas of natural and human spheres. This study outlines the background within which these techniques developed, the fundamental elements on which they are based and details some of the predictive techniques. This study aims to provide some insight into their mechanisms and their potential.