Applications of Wavelet Neural Networks in Time Series Forecasting  (Papaerback, Mohd Yasin Pir)

The aim is to presents a text that can be helpful to teach wavelet networks in an organized manner. This book can serve as a valuable reference for researchers and learners in the fields of mathematical modelling, artificial neural networks, wavelet analysis and economics. This book provides both theoretical and practical applications of wavelet networks. The book is rationally divided into three parts. The first part presents the foundation aspects of wavelet neural networks and the second part presents the applications of wavelet neural networks in the areas of finance and classification. The third part presents the future direction and conclusion in concise manner. There are four chapters in first part. Chapter I presents introduction of time series analysis. The main objective of time series analysis is to describe the data using summary statistics and to develop a suitable forecasting model for estimating future values. Chapter 2 introduces Artificial Neural Networks and Wavelet Analysis. Artificial Neural Networks can generalize any non-linear relationship and as such have become very useful tool for time series forecasting for few decades now. They do not require any earlier assumption about the underlying relationship for a particular problem. On the other hand, Wavelet analysis has an advantage of capturing features in the time series that vary across both time and frequency domains. It can handle very irregular data series with applications that range from short term forecasting to the testing of market models and have been found very successful in forecasting time series data from the last two decades. Chapter 3 briefly introduces various statistical methods such as Exponential Smoothing, Regression, ARIMA, Threshold and GARCH etc that are used for forecasting future values. It also emphasis the use of artificial neural networks and wavelets as preferred methods of forecasting. Chapter 4 introduces wavelet neural networks in financial time series forecasting. It proposes the use of hybrid model of wavelets and neural networks to forecast time series values. The second part is divided into two parts. Chapter 6 introduces few applications of wavelet networks in the areas of finance and classification with the help of experimental results and discussions. Chapter 7 presents the conclusion of the whole study and future direction in a concise manner.