Damped trend exponential smoothing: A modelling viewpoint

Edward Mckenzie, Everette S. Gardner Jr

Research output: Contribution to journalArticle

21 Citations (Scopus)

Abstract

Over the past twenty years, damped trend exponential smoothing has performed well in numerous empirical studies, and it is now well established as an accurate forecasting method. The original motivation for this method was intuitively appealing, but said very little about why or when it provided an optimal approach. The aim of this paper is to provide a theoretical rationale for the damped trend method based on Brown's original thinking about the form of underlying models for exponential smoothing. We develop a random coefficient state space model for which damped trend smoothing provides an optimal approach, and within which the damping parameter can be interpreted directly as a measure of the persistence of the linear trend.
LanguageEnglish
Pages661-665
Number of pages5
JournalInternational Journal of Forecasting
Volume26
Issue number4
DOIs
Publication statusPublished - Dec 2010

Fingerprint

Modeling
Exponential smoothing
Persistence
Random coefficients
Empirical study
Forecasting method
State-space model
Rationale
Smoothing

Keywords

  • time series
  • state space models
  • ARIMA models
  • exponential smoothing

Cite this

@article{aae57420eeb14c319f36aacfbd0497b1,
title = "Damped trend exponential smoothing: A modelling viewpoint",
abstract = "Over the past twenty years, damped trend exponential smoothing has performed well in numerous empirical studies, and it is now well established as an accurate forecasting method. The original motivation for this method was intuitively appealing, but said very little about why or when it provided an optimal approach. The aim of this paper is to provide a theoretical rationale for the damped trend method based on Brown's original thinking about the form of underlying models for exponential smoothing. We develop a random coefficient state space model for which damped trend smoothing provides an optimal approach, and within which the damping parameter can be interpreted directly as a measure of the persistence of the linear trend.",
keywords = "time series, state space models, ARIMA models, exponential smoothing",
author = "Edward Mckenzie and {Gardner Jr}, {Everette S.}",
year = "2010",
month = "12",
doi = "10.1016/j.ijforecast.2009.07.001",
language = "English",
volume = "26",
pages = "661--665",
journal = "International Journal of Forecasting",
issn = "0169-2070",
number = "4",

}

Damped trend exponential smoothing : A modelling viewpoint. / Mckenzie, Edward; Gardner Jr, Everette S.

In: International Journal of Forecasting, Vol. 26, No. 4, 12.2010, p. 661-665.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Damped trend exponential smoothing

T2 - International Journal of Forecasting

AU - Mckenzie, Edward

AU - Gardner Jr, Everette S.

PY - 2010/12

Y1 - 2010/12

N2 - Over the past twenty years, damped trend exponential smoothing has performed well in numerous empirical studies, and it is now well established as an accurate forecasting method. The original motivation for this method was intuitively appealing, but said very little about why or when it provided an optimal approach. The aim of this paper is to provide a theoretical rationale for the damped trend method based on Brown's original thinking about the form of underlying models for exponential smoothing. We develop a random coefficient state space model for which damped trend smoothing provides an optimal approach, and within which the damping parameter can be interpreted directly as a measure of the persistence of the linear trend.

AB - Over the past twenty years, damped trend exponential smoothing has performed well in numerous empirical studies, and it is now well established as an accurate forecasting method. The original motivation for this method was intuitively appealing, but said very little about why or when it provided an optimal approach. The aim of this paper is to provide a theoretical rationale for the damped trend method based on Brown's original thinking about the form of underlying models for exponential smoothing. We develop a random coefficient state space model for which damped trend smoothing provides an optimal approach, and within which the damping parameter can be interpreted directly as a measure of the persistence of the linear trend.

KW - time series

KW - state space models

KW - ARIMA models

KW - exponential smoothing

UR - http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V92-4WXBTM9-1&_user=875629&_coverDate=12%2F31%2F2010&_rdoc=1&_fmt=high&_orig=gateway&_origin=gateway&_sort=d&_docanchor=&view=c&_searchStrId=1700853985&_rerunOrigin=google&_acct=C000046979&_version=1&_urlVersion=0&_userid=875629&md5=8f9d51baf0569bddee1a072d6e564328&searchtype=a

U2 - 10.1016/j.ijforecast.2009.07.001

DO - 10.1016/j.ijforecast.2009.07.001

M3 - Article

VL - 26

SP - 661

EP - 665

JO - International Journal of Forecasting

JF - International Journal of Forecasting

SN - 0169-2070

IS - 4

ER -