Scribd
Upload
27 views9 pages

Tong 1983

Uploaded by

katlina lin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
27 views9 pages

Tong 1983

Uploaded by

katlina lin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

Lecture Notes in Statistics

Vol. 1: R. A. Fisher: An Appreciation. Edited by S. E. Fienberg and D. V. Hinkley. xi, 208


pages, 1980.

Vol. 2: Mathematical Statistics and Probability Theory. Proceedings 1978. Edited by W.


Klonecki, A. Kozek, and J. Rosinski. xxiv, 373 pages, 1980.

Vol. 3: B. D. Spencer, Benefit-Cost Analysis of Data Used to Allocate Funds. viii, 296
pages, 1980.

Vol. 4: E. A. van Doorn, Stochastic Monotonicity and Queueing Applications of Birth-


Death Processes. vi, 118 pages, 1981.

Vol. 5: T. Rolski, Stationary Random Processes Associated with Point Processes. vi,
139 pages, 1981.

Vol. 6: S. S. Gupta and D.-Y. Huang, Multiple Statistical Decision Theory: Recent
Developments. viii, 104 pages, 1981.

Vol. 7: M. Akahira and K. Takeuchi, Asymptotic Efficiency of Statistical Estimators. viii,


242 pages, 1981.

Vol. 8: The First Pannonian Symposium on Mathematical Statistics. Edited by P.


Revesz, L. Schmetterer, and V. M. Zolotarev. vi, 308 pages, 1981.

Vol. 9: B. J¢rgensen, Statistical Properties of the Generalized Inverse Gaussian Distri-


bution. vi, 188 pages, 1981.

Vol. 10: A. A. Mcintosh, Fitting Linear Models: An Application of Conjugate Gradient


Algorithms. vi, 200 pages, 1982.

Vol. 11: D. F. Nicholls and B. G. Quinn, Random Coefficient Autoregressive Models: An


Introduction. v, 154 pages, 1982.

Vol. 12: M. Jacobsen, Statistical Analysis of Counting Processes. vii, 226 pages, 1982.

Vol. 13: J. Pfanzagl (with the assistance of W. Wefelmeyer), Contributions to a General


Asymptotic Statistical Theory. vii, 315 pages, 1982.

Vol. 14: GUM 82: Proceedings of the International Conference on Generalised Linear
Models. Edited by R. Gilchrist. v, 188 pages, 1982.

Vol. 15: K. R. W. Brewer and M. Hanif, Sampling with Unequal Probabilities. ix, 164
pages, 1983.

Vol. 16: Specifying Statistical Models: From Parametric to Non-Parametric, Using


Bayesian or Non-Bayesian Approaches. Edited by J. P. Florens, M. Mouchart, J. P.
Raoult, L. Simar, and A. F. M. Smith. xi, 204 pages, 1983.

Vol. 17: I. V. Basawa and D. J. Scott, Asymptotic Optimal Inference for Non-Ergodic
Models. ix, 170 pages, 1983.
Lecture Notes in Statistics

Vol. 18: W. Britton, Conjugate Duality and the Exponential Fourier Spectrum. v, 226
pages, 1983.

Vol. 19: L. Fernholz, von Mises Calculus for Statistical Functionals. viii, 124 pages,
1983.

Vol. 20: Mathematical Learning Models- Theory and Algororithms: Proceedings of a


Conference. Edited by U. Herkenrath, D. Kalin, W. Vogel. xiv, 226 pages, 1983.

Vol. 21: H. Tong, Threshold Models in Non-linear Time Series Analysis. x, 323 pages,
1983.
Lecture Notes in
Statistics
Edited by D. Brillinger, S. Fienberg, J. Gani,
J. Hartigan, and K. Krickeberg

21

Howell Tong

Threshold Models in Non-linear


Time Series Analysis

Springer-Verlag
New York Berlin Heidelberg Tokyo
Howell Tong
The Chinese University
of Hong Kong
Department of Statistics
Shatin, NT
Hong Kong

AMS Subject Classification: 62MIO, 62M15

Library of Congress Cataloging in Publication Data


Tong, Howell.
Threshold models in non-linear time series analysis
(Lecture notes in statistics; 21)
Bibliography: p.
Includes indexes.
1. Time-series analysis. 2. Spectral theory
(Mathematics) I. Title. II. Series: Lecture notes in
statistics (Springer-Verlag); v. 21.
QA280.T6 1983 519.5'5 83-17026

With 154 illustrations.

© 1983 by Springer-Verlag New York Inc.


Al! rights reserved. No part of this book may be translated or reproduced in any
form without written permission from Springer-Verlag, 175 Fifth Avenue,
New York, New York, 10010, U.S.A.

9 876 54 32 I

ISBN- 13: 978-0-387-90918-9 e-ISBN- 13: 978-1-4684-7888-4


001: 10.1007/978-1-4684-7888-4
To MMY, S-imOYl aYld AYlYla,

-- My TlvtuhoU6.
PREFACE

In the last two years or so, I was most fortunate in being given
opportunities of lecturing on a new methodology to a variety of audiences
in Britain, China, Finland, France and Spain. Despite my almost Confucian
attitude of preferring talking (i.e. a transient record) to writing (i.e.
a permanent record), the warm encouragement of friends has led to the
ensuing notes. I am also only too conscious of the infancy of the
methodology introduced in these notes. However, it is my sincere hope
that exposure to a wider audience will accelerate its maturity. Readers
are assumed to be familiar with the basic theory of time series analysis.
The book by Professor M.B. Priestley (1981) may be used as a general
reference.

Chapter One is addressed to the general question: "why do we need


non-linear time series models?" After describing some significant advantages
of linear models, it singles out several major limitations of linearity.
Of course, the selection reflects my personal view on the subject, which
is only at its very beginning, although there does seem to be a general
agreement in the literature that time irr'eversibility and limit cycles
are among the most obvious.

Since the non-linear world is so vast, the first vital question must
be concerned with which mode of non-linearity to be studied. Chapter Two
is devoted to this question of 'which'. The subject matter has a very
strong physical flavour and the concept of a limit cycle is given a
prominent position. I have given some of the mathematical and physical
details in connection with this concept. It is hoped that they will help
readers new to this field to appreciate its beauty and usefulness for the
modelling of cyclical data. Numerous examples are descrilJed which are
drawn from diverse fields; they include electronics, oceanography, hydrology,
ecology, marine engineering, medical engineering, solar astrophysics, etc.
The connecting theme which emerges is piece-wise linearity.

Chapter Three develops the idea of piece-wise linearization for time


series modelling, in a systematic way. A full description of the class of
threshold time series models is given. Probabilistic and structural aspects
of these models are described in detail. These include ergodicity,
stationary distribution, moments and non-linear vibrations. Some discussion
about the generality of the newmode1s is also included. Multi-step-ahead

vi i
vi Ii

forecasting using these models is also illustrated.

Chapter Four describes, in full detail, the identification of


threshold models. A fairly extensive discussion of Akaike's entropy
maximization principle is given ab initio. The significant
importance of diagnostics is emphasized in these notes. In fact, three
major diagnostics are proposed. Sampling properties of the parameter
estimates and data transformation are also included.

Chapter Five gives a thorough account of several case studies. They


have been chosen variously from ecology, solar astrophysics and hydrology.
The motivation behind these case studies is to emphasize the view that
non-linear time series models have provided us with a valuable opportunity
of linking time series analysis more closely with the background physics,
biology, etc. The link can only be of mutual benefit to both statistics
itself and the scientific community the former is supposed to serve. It
is my hope that the fitted models described in this Chapter do represent
a modest step closer to real i ty. In any case, I have tri ed and it \~as
surely fun in building them.

The infant was conceived in Manchester and delivered i~ Hong Kong.


a~ therefore most grateful to my former colleagues and postgraruate
students at the UniverSity of Manchester Institute of Science &Technology,
who together constituted the necessary environment for the conception.
In particular, it is a pleasure for me to express my sincere gratitude to
Professor Maurice Priestley, from whom I learnt my first lessons of linear
models and obtained encouragement when I was trying to bend them. My
appreciation goes to Professor David Brillinger, who has made many valuable
criticisms and suggestions on the manuscript. I also express my sincere
thanks to Mrs. Clare Rowson and Mr. Billy Lam for their efficient typing
and assistance in the general preparations of the manuscript for camera-
readiness. The appended computer programs are the results of enhancement
and refinement by Dr. K.S. Lilli, Messrs. J. Pemberton, Z.M. Wu, B. Thanoon,
D. Nield and K.S. Chan; I remain indebted to all of them.

H. TONG

Shatin, Hong Kong


May 1983
CONTENTS

CHAPTER ONE INTRODUCTION Page


1. TIME SERIES MODEL BUILDING 1
2. STATIONARITY 3
3. LINEAR GAUSSIAN MODELS 6
4. SOME ADVANTAGES AND SOME LIMITATIONS OF ARMA MODELS B
5. WHAT NEXT? 31

CHAPTER TWO SOME BASIC CONCEPTS


1. ORI ENTATION 34
2. LIMIT CYCLES 35
3. SOME EXA~lPLES OF THRESHOLD MODELS 44
4. TIME DELAY 54
5. DISCUSSION 58

CHAPTER THREE THRESHOLD MODELS


1. A CANONICAL FORM 59
2. GENERALITY OF SETAR MODELS 64
3. NON-LINEAR DIFFERENCE EQUATIONS 71
4. THRESHOLD MODELS AND DISCRETE-TIME NON-LINEAR VIBRATIONS 77
5. ERGODICITY 93
6. STATIONARY DISTRIBUTIONS AND ,MOMENTS 95
7. CYCLICAL STRUCTURE AND MULTI-STEP-AHEAD FORECASTING 107

CHAPTER FOUR IDENTIFICATION


1. A GENERAL PRINCIPLE 122
2. ESTIMATION OF PARAMETERS 131
3. SAMPLING PROPERTIES 136
3.1 A General Result 137
3.2 Application of General Result to SETAR 139
3.3 Some Simulation Studies 141
3.4 Standard Errors of Some Parameter Estimates 147
3.5 Unsolved Problems 149
4. DIAGNOSTICS AND GRAPHICAL METHODS 150
5. MISCELLANEA 160

CHAPTER FIVE SOME CASE STUDIES


1. ANALYSIS OF SOME ECOLOGICAL DATA 163
1.1 Ecological Background 163

ix
x

Page
1.2 Some Graphical Results 165
1.3 A Full Identification 185
1.4 Diagnostics 190
1.5 Transformation 210
1.6 Extension to Two Species 214
2. ANALYSIS OF THE SUNSPOT NUMBERS 230
2.1 Some Background 230
2.2 SETAR Modelling 231
2.3 Transformation 248
2.4 Multi-step-ahead Forecasting 252
2.5 Some Discussion 252
3. ANALYSIS OF SOME RIVERFLOW DATA 257
3.1 Some Background 257
3.2 TAR SO Modelling of Vatnsdalsa River, Iceland (1972-1974) 260
3.3 TARSO Modelling of Jokulsa River, Iceland (1972-1974) 266
3.4 Some Discussion 271
4. A CASE STUDY WITH LABORATORY DATA 272
5. A FUZZY EXTENSION 276
6. CONCLUDING REMARKS 279

APPEDICES 280

REFERENCES 307

AUTHOR INDEX 317


SUBJECT INDEX 321

You might also like