A unified approach to estimating and modeling linear and nonlinear time series

Cathy W S Chen, Robert McCulloch, Ruey S. Tsay

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

In this article, we propose a unified approach to estimating and modeling univariate time series. The approach applies to both linear and nonlinear time series models and can be used to discriminate non-nested nonlinear models. For example, it can discriminate between threshold autoregressive and bilinear models or between autoregressive and moving average models. It can also be used to estimate and discriminate between symmetric and asymmetric conditional heteroscedastic models commonly used in volatility studies of financial time series. The proposed approach is based on Gibbs sampling and may require substantial amounts of computing in some applications. We illustrate the proposed approach by some simulated and real examples. Comparison with other existing methods is also discussed.

Original languageEnglish (US)
Pages (from-to)451-472
Number of pages22
JournalStatistica Sinica
Volume7
Issue number2
StatePublished - Apr 1997
Externally publishedYes

Fingerprint

Threshold Autoregressive Model
Nonlinear Time Series Model
Bilinear Model
Heteroscedastic Model
Moving Average Model
Nonlinear Time Series
Conditional Model
Financial Time Series
Gibbs Sampling
Volatility
Univariate
Nonlinear Model
Time series
Computing
Modeling
Estimate
Nonlinear time series

Keywords

  • Bayesian model selection
  • Bilinear model
  • Gibbs sampler
  • Mixed model
  • Stochastic volatility
  • Threshold autoregressive model

ASJC Scopus subject areas

  • Mathematics(all)
  • Statistics and Probability

Cite this

A unified approach to estimating and modeling linear and nonlinear time series. / Chen, Cathy W S; McCulloch, Robert; Tsay, Ruey S.

In: Statistica Sinica, Vol. 7, No. 2, 04.1997, p. 451-472.

Research output: Contribution to journalArticle

@article{5d3a98d8082041a99e2533a8905f9e30,
title = "A unified approach to estimating and modeling linear and nonlinear time series",
abstract = "In this article, we propose a unified approach to estimating and modeling univariate time series. The approach applies to both linear and nonlinear time series models and can be used to discriminate non-nested nonlinear models. For example, it can discriminate between threshold autoregressive and bilinear models or between autoregressive and moving average models. It can also be used to estimate and discriminate between symmetric and asymmetric conditional heteroscedastic models commonly used in volatility studies of financial time series. The proposed approach is based on Gibbs sampling and may require substantial amounts of computing in some applications. We illustrate the proposed approach by some simulated and real examples. Comparison with other existing methods is also discussed.",
keywords = "Bayesian model selection, Bilinear model, Gibbs sampler, Mixed model, Stochastic volatility, Threshold autoregressive model",
author = "Chen, {Cathy W S} and Robert McCulloch and Tsay, {Ruey S.}",
year = "1997",
month = "4",
language = "English (US)",
volume = "7",
pages = "451--472",
journal = "Statistica Sinica",
issn = "1017-0405",
publisher = "Institute of Statistical Science",
number = "2",

}

TY - JOUR

T1 - A unified approach to estimating and modeling linear and nonlinear time series

AU - Chen, Cathy W S

AU - McCulloch, Robert

AU - Tsay, Ruey S.

PY - 1997/4

Y1 - 1997/4

N2 - In this article, we propose a unified approach to estimating and modeling univariate time series. The approach applies to both linear and nonlinear time series models and can be used to discriminate non-nested nonlinear models. For example, it can discriminate between threshold autoregressive and bilinear models or between autoregressive and moving average models. It can also be used to estimate and discriminate between symmetric and asymmetric conditional heteroscedastic models commonly used in volatility studies of financial time series. The proposed approach is based on Gibbs sampling and may require substantial amounts of computing in some applications. We illustrate the proposed approach by some simulated and real examples. Comparison with other existing methods is also discussed.

AB - In this article, we propose a unified approach to estimating and modeling univariate time series. The approach applies to both linear and nonlinear time series models and can be used to discriminate non-nested nonlinear models. For example, it can discriminate between threshold autoregressive and bilinear models or between autoregressive and moving average models. It can also be used to estimate and discriminate between symmetric and asymmetric conditional heteroscedastic models commonly used in volatility studies of financial time series. The proposed approach is based on Gibbs sampling and may require substantial amounts of computing in some applications. We illustrate the proposed approach by some simulated and real examples. Comparison with other existing methods is also discussed.

KW - Bayesian model selection

KW - Bilinear model

KW - Gibbs sampler

KW - Mixed model

KW - Stochastic volatility

KW - Threshold autoregressive model

UR - http://www.scopus.com/inward/record.url?scp=0031529932&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0031529932&partnerID=8YFLogxK

M3 - Article

VL - 7

SP - 451

EP - 472

JO - Statistica Sinica

JF - Statistica Sinica

SN - 1017-0405

IS - 2

ER -