A note on the analysis of short AR(1) type time series models with replicated observations

Ponnuthurai Ainkaran; Robert Mellor; V. Pemajayantha; Robert Mellor; M. Shelton Peiris; Jay R. Rajasekera

A note on the analysis of short AR(1) type time series models with replicated observations

Ponnuthurai Ainkaran
, Robert Mellor
, V. Pemajayantha
, Robert Mellor
, M. Shelton Peiris
, Jay R. Rajasekera

International University of Japan

Research output: Chapter in Book / Conference Paper › Conference Paper

Abstract

Analysis of a large number of independent replications from short, AR(1) type time series is considered. Maximum likelihood estimation (mle) procedures based on both conditional and exact arguments are discussed. A simulation study is carried out to examine the finite sample behavior of the estimates. We compare the bias and the mean square error (mse) of these estimates with their corresponding asymptotic results. It is shown that the bias and the mse are very small for large number of replicates.

Original language	English
Title of host publication	Current Research in Modelling, Data Mining & Quantitative Techniques
Publisher	University of Western Sydney
Number of pages	14
ISBN (Electronic)	0975159909
ISBN (Print)	9780975159903
Publication status	Published - 2003
Event	Workshop on Advanced Research Methods - Duration: 1 Jan 2003 → …

Conference

Conference	Workshop on Advanced Research Methods
Period	1/01/03 → …

Keywords

time-series analysis
correlation (statistics)
estimation theory
error analysis (mathematics)
measurement
numerical analysis

Cite this

@inproceedings{ded4bc2bad404291a6966690a59f4fe4,

title = "A note on the analysis of short AR(1) type time series models with replicated observations",

abstract = "Analysis of a large number of independent replications from short, AR(1) type time series is considered. Maximum likelihood estimation (mle) procedures based on both conditional and exact arguments are discussed. A simulation study is carried out to examine the finite sample behavior of the estimates. We compare the bias and the mean square error (mse) of these estimates with their corresponding asymptotic results. It is shown that the bias and the mse are very small for large number of replicates.",

keywords = "time-series analysis, correlation (statistics), estimation theory, error analysis (mathematics), measurement, numerical analysis",

author = "Ponnuthurai Ainkaran and Robert Mellor and V. Pemajayantha and Robert Mellor and Peiris, \{M. Shelton\} and Rajasekera, \{Jay R.\}",

year = "2003",

language = "English",

isbn = "9780975159903",

booktitle = "Current Research in Modelling, Data Mining \& Quantitative Techniques",

publisher = "University of Western Sydney",

note = "Workshop on Advanced Research Methods ; Conference date: 01-01-2003",

}

TY - GEN

T1 - A note on the analysis of short AR(1) type time series models with replicated observations

AU - Ainkaran, Ponnuthurai

AU - Mellor, Robert

AU - Pemajayantha, V.

AU - Mellor, Robert

AU - Peiris, M. Shelton

AU - Rajasekera, Jay R.

PY - 2003

Y1 - 2003

N2 - Analysis of a large number of independent replications from short, AR(1) type time series is considered. Maximum likelihood estimation (mle) procedures based on both conditional and exact arguments are discussed. A simulation study is carried out to examine the finite sample behavior of the estimates. We compare the bias and the mean square error (mse) of these estimates with their corresponding asymptotic results. It is shown that the bias and the mse are very small for large number of replicates.

AB - Analysis of a large number of independent replications from short, AR(1) type time series is considered. Maximum likelihood estimation (mle) procedures based on both conditional and exact arguments are discussed. A simulation study is carried out to examine the finite sample behavior of the estimates. We compare the bias and the mean square error (mse) of these estimates with their corresponding asymptotic results. It is shown that the bias and the mse are very small for large number of replicates.

KW - time-series analysis

KW - correlation (statistics)

KW - estimation theory

KW - error analysis (mathematics)

KW - measurement

KW - numerical analysis

UR - http://handle.uws.edu.au:8081/1959.7/489318

M3 - Conference Paper

SN - 9780975159903

BT - Current Research in Modelling, Data Mining & Quantitative Techniques

PB - University of Western Sydney

T2 - Workshop on Advanced Research Methods

Y2 - 1 January 2003

ER -

A note on the analysis of short AR(1) type time series models with replicated observations

Abstract

Conference

Keywords

Other files and links

Fingerprint

Cite this