Selecting the forgetting factor in subset autoregressive modelling

T. J. Brailsford, Jack H W Penm, R. D. Terrell

Research output: Contribution to journalArticleResearchpeer-review

14 Citations (Scopus)

Abstract

Conventional methods to determine the forgetting factors in autoregressive (AR) models are mostly based on arbitrary or personal choices. In this paper, we present two procedures which can be used to select the forgetting factor in subset AR modelling. The first procedure uses the bootstrap to determine the value of a fixed forgetting factor. The second procedure starts from this base and applies the time-recursive maximum likelihood estimation to a variable forgetting factor. In one illustration using real exchange rates, we demonstrate the effect of the forgetting factor in subset AR modelling on ex ante forecasting of non-stationary time series. In a second illustration, these two procedures are applied to time-update forecasts for a stock market index. Subset AR models not including a forgetting factor act as a set of benchmarks for assessing ex ante forecasting performance, and consistently improved forecasting performance is demonstrated for these proposed procedures.

Original languageEnglish
Pages (from-to)629-649
Number of pages21
JournalJournal of Time Series Analysis
Volume23
Issue number6
DOIs
Publication statusPublished - 2002
Externally publishedYes

Fingerprint

Subset
Modeling
Forecasting
Maximum likelihood estimation
Autoregressive Model
Set theory
Time series
Real Exchange Rate
Recursive Estimation
Non-stationary Time Series
Stock Market
Maximum Likelihood Estimation
Bootstrap
Forecast
Factors
Forgetting
Update
Benchmark
Arbitrary
Demonstrate

Cite this

Brailsford, T. J. ; Penm, Jack H W ; Terrell, R. D. / Selecting the forgetting factor in subset autoregressive modelling. In: Journal of Time Series Analysis. 2002 ; Vol. 23, No. 6. pp. 629-649.
@article{d96bfeae8a674801808ba4c3dec2b5b6,
title = "Selecting the forgetting factor in subset autoregressive modelling",
abstract = "Conventional methods to determine the forgetting factors in autoregressive (AR) models are mostly based on arbitrary or personal choices. In this paper, we present two procedures which can be used to select the forgetting factor in subset AR modelling. The first procedure uses the bootstrap to determine the value of a fixed forgetting factor. The second procedure starts from this base and applies the time-recursive maximum likelihood estimation to a variable forgetting factor. In one illustration using real exchange rates, we demonstrate the effect of the forgetting factor in subset AR modelling on ex ante forecasting of non-stationary time series. In a second illustration, these two procedures are applied to time-update forecasts for a stock market index. Subset AR models not including a forgetting factor act as a set of benchmarks for assessing ex ante forecasting performance, and consistently improved forecasting performance is demonstrated for these proposed procedures.",
author = "Brailsford, {T. J.} and Penm, {Jack H W} and Terrell, {R. D.}",
year = "2002",
doi = "10.1111/1467-9892.00283",
language = "English",
volume = "23",
pages = "629--649",
journal = "Journal of Time Series Analysis",
issn = "0143-9782",
publisher = "Wiley-Blackwell",
number = "6",

}

Selecting the forgetting factor in subset autoregressive modelling. / Brailsford, T. J.; Penm, Jack H W; Terrell, R. D.

In: Journal of Time Series Analysis, Vol. 23, No. 6, 2002, p. 629-649.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Selecting the forgetting factor in subset autoregressive modelling

AU - Brailsford, T. J.

AU - Penm, Jack H W

AU - Terrell, R. D.

PY - 2002

Y1 - 2002

N2 - Conventional methods to determine the forgetting factors in autoregressive (AR) models are mostly based on arbitrary or personal choices. In this paper, we present two procedures which can be used to select the forgetting factor in subset AR modelling. The first procedure uses the bootstrap to determine the value of a fixed forgetting factor. The second procedure starts from this base and applies the time-recursive maximum likelihood estimation to a variable forgetting factor. In one illustration using real exchange rates, we demonstrate the effect of the forgetting factor in subset AR modelling on ex ante forecasting of non-stationary time series. In a second illustration, these two procedures are applied to time-update forecasts for a stock market index. Subset AR models not including a forgetting factor act as a set of benchmarks for assessing ex ante forecasting performance, and consistently improved forecasting performance is demonstrated for these proposed procedures.

AB - Conventional methods to determine the forgetting factors in autoregressive (AR) models are mostly based on arbitrary or personal choices. In this paper, we present two procedures which can be used to select the forgetting factor in subset AR modelling. The first procedure uses the bootstrap to determine the value of a fixed forgetting factor. The second procedure starts from this base and applies the time-recursive maximum likelihood estimation to a variable forgetting factor. In one illustration using real exchange rates, we demonstrate the effect of the forgetting factor in subset AR modelling on ex ante forecasting of non-stationary time series. In a second illustration, these two procedures are applied to time-update forecasts for a stock market index. Subset AR models not including a forgetting factor act as a set of benchmarks for assessing ex ante forecasting performance, and consistently improved forecasting performance is demonstrated for these proposed procedures.

UR - http://www.scopus.com/inward/record.url?scp=0040078345&partnerID=8YFLogxK

U2 - 10.1111/1467-9892.00283

DO - 10.1111/1467-9892.00283

M3 - Article

VL - 23

SP - 629

EP - 649

JO - Journal of Time Series Analysis

JF - Journal of Time Series Analysis

SN - 0143-9782

IS - 6

ER -