Editor: Cagdas H. Aladag

Advances in Time Series Forecasting

Volume 1

eBook: US $49 Special Offer (PDF + Printed Copy): US $113
Printed Copy: US $89
Library License: US $196
ISSN: 2543-2796 (Print)
ISSN: 2543-280X (Online)
ISBN: 978-1-60805-522-7 (Print)
ISBN: 978-1-60805-373-5 (Online)
Year of Publication: 2012
DOI: 10.2174/97816080537351120101

Introduction

Time series analysis is applicable in a variety of disciplines, such as business administration, economics, public finance, engineering, statistics, econometrics, mathematics and actuarial sciences. Forecasting the future assists in critical organizational planning activities. Time series analysis is employed by many different organizations such as hospitals, universities, commercial enterprises or government organizations in order to forecast future scenarios. Therefore, many time series forecasting methods have been proposed and improved in statistical literature. Linear models such as Box-Jenkins methods were earlier used in many situations. Then, to overcome the restrictions of these linear models and to account for certain nonlinear patterns observed in real problems, some nonlinear models are also presented in literature. However, since these nonlinear models were developed for specific nonlinear patterns, they are not suitable for modeling other types of nonlinearity in time series. In recent years, efficient and advanced techniques such as artificial neural networks, fuzzy time series and some hybrid models have been used to forecast any kind of real life time series analyses. Both theoretical and empirical findings in academic literature show that these approaches give comparatively reliable forecasts than those obtained from conventional forecasting methods. In addition, conventional models require some assumptions such as linearity and normal distribution or cannot be utilized efficiently for some real time series such as temperature and share prices of stockholders since these kind of series contain some uncertainty. However, when advanced methods such as neural networks and fuzzy time series are used to forecast time series, there is no need to satisfy any assumption and the time series containing uncertainty can be forecasted efficiently

This e-book contains recent effective applications and descriptions of these advanced forecasting methods. Readers will learn how these methods work and how these approaches can be used to forecast real life time series. In addition, the hybrid forecasting model approach, which combines different methods to obtain better forecast results, is also explained. Readers can also find the applications of hybrid forecasting models in this e-book. This e-book also enables skilled statisticians to create a new hybrid forecasting model suitable for their own objectives. Data presented in this e-book is problem based and is taken from real life situations. This e-book is a valuable resource for students, statisticians and working professionals interested in advanced time series analysis.

Preface

Time series analysis has got attention of many researches from different fields, such as business administration, economics, public finances, engineering, statistics, econometrics, mathematics and actuarial sciences. When many organizations are planning their future, they have to forecast the future. Time series analysis has been employed by many organizations, such as hospitals, universities, companies or government organizations in order to forecast how could be the future. Therefore, many time series forecasting methods have been proposed and improved in the literature. Firstly, linear models such as Box-Jenkins methods were used in many areas of time series forecasting. Furthermore, to overcome the restriction of the linear models and to account for certain nonlinear patterns observed in real problems, some nonlinear models have been proposed in the literature. However, since these nonlinear models were developed for specific nonlinear patterns, they are not capable of modeling other types of nonlinearity in time series. In recent years, to overcome these issues, efficient soft computing techniques such as artificial neural networks, fuzzy time series and some hybrid models have been used to forecast any kind of real life time series. Both theoretical and empirical findings in the literature show that these approaches give better forecasts than those obtained from conventional forecasting methods. In addition, conventional models require some assumptions such as linearity and normal distribution cannot be utilized efficiently for some real time series such as temperature and share prices of stockholders, since this kind of series contain some uncertainty in itself. However, when soft computing methods such as neural networks and fuzzy time series are used to forecast time series, there is no need to satisfy any assumption and the time series uncertainty can be forecasted efficiently.

This eBook contains recent applications and descriptions of these effective soft computing methods. The readers can learn how these methods work and how these approaches can be used to forecast real life time series. In addition, the hybrid forecasting model approach, which is based on combining different soft computing methods to get better forecasts, is explained and at the same time, the reader can find the applications of hybrid forecasting models. The reader of this eBook can also create a new hybrid forecasting model. Although the soft computing forecasting models have many advantages, at the same time there are still some problems with their usage. These problems are pointed out in this ebook. After researchers see those problems, they make some contributions to these forecasting methods by filling some gaps to obtain better forecast results. Furthermore, some new forecasting models are introduced in the eBook.

Cagdas Hakan Aladag
Department of Statistics
Hacettepe University
Bolzano University
Turkey

RELATED BOOKS

.Reliability Calculations with the Stochastic Finite Element.
.Recent Advances in Lifetime and Reliability Models.
.Advances in Time Series Forecasting.