A Study on the nature of an additive outlier in ARMA(1,1) models

Azami Zaharim, Rafizah Rajali, Raden Mohamad Atok, Kamarulzaman Ibrahim, Ahmad Mahir Razali

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Outliers are often encountered in time series analysis. Types of outliers that are usually dealt with are additive outlier, innovational outlier, temporary change and level shift. This study focuses solely on the additive outlier as it is the most common type found in time series due to its association with human error such as typing and recording mistakes. To understand the nature of an additive outlier, simulations of ARMA (1, 1) time series are contaminated with an additive outlier. Then, we examine the effect of the respected additive outlier on observations and residuals. It is found that an additive outlier does not affect observations prior to its existence and the ones subsequent to it. However, given the occurrence of the additive outlier at t=T, the additive outlier is seen to have a noticeable effect on the residual of t=T and t=T+1.

Original languageEnglish
Pages (from-to)362-368
Number of pages7
JournalEuropean Journal of Scientific Research
Volume32
Issue number3
Publication statusPublished - 2009

Fingerprint

Additive Outliers
Autoregressive Moving Average
outlier
Time series
time series analysis
Outlier
Time series analysis
Model
Level Shift
Human Error
additive effect
Time Series Analysis
time series

Keywords

  • ARMA (1
  • Simulation Study
  • The Nature of Additive Outlier

ASJC Scopus subject areas

  • General

Cite this

A Study on the nature of an additive outlier in ARMA(1,1) models. / Zaharim, Azami; Rajali, Rafizah; Atok, Raden Mohamad; Ibrahim, Kamarulzaman; Razali, Ahmad Mahir.

In: European Journal of Scientific Research, Vol. 32, No. 3, 2009, p. 362-368.

Research output: Contribution to journalArticle

Zaharim, Azami ; Rajali, Rafizah ; Atok, Raden Mohamad ; Ibrahim, Kamarulzaman ; Razali, Ahmad Mahir. / A Study on the nature of an additive outlier in ARMA(1,1) models. In: European Journal of Scientific Research. 2009 ; Vol. 32, No. 3. pp. 362-368.
@article{57f6674809164bf48f9e4ffeca0830f4,
title = "A Study on the nature of an additive outlier in ARMA(1,1) models",
abstract = "Outliers are often encountered in time series analysis. Types of outliers that are usually dealt with are additive outlier, innovational outlier, temporary change and level shift. This study focuses solely on the additive outlier as it is the most common type found in time series due to its association with human error such as typing and recording mistakes. To understand the nature of an additive outlier, simulations of ARMA (1, 1) time series are contaminated with an additive outlier. Then, we examine the effect of the respected additive outlier on observations and residuals. It is found that an additive outlier does not affect observations prior to its existence and the ones subsequent to it. However, given the occurrence of the additive outlier at t=T, the additive outlier is seen to have a noticeable effect on the residual of t=T and t=T+1.",
keywords = "ARMA (1, Simulation Study, The Nature of Additive Outlier",
author = "Azami Zaharim and Rafizah Rajali and Atok, {Raden Mohamad} and Kamarulzaman Ibrahim and Razali, {Ahmad Mahir}",
year = "2009",
language = "English",
volume = "32",
pages = "362--368",
journal = "European Journal of Scientific Research",
issn = "1450-202X",
publisher = "European Journals Inc.",
number = "3",

}

TY - JOUR

T1 - A Study on the nature of an additive outlier in ARMA(1,1) models

AU - Zaharim, Azami

AU - Rajali, Rafizah

AU - Atok, Raden Mohamad

AU - Ibrahim, Kamarulzaman

AU - Razali, Ahmad Mahir

PY - 2009

Y1 - 2009

N2 - Outliers are often encountered in time series analysis. Types of outliers that are usually dealt with are additive outlier, innovational outlier, temporary change and level shift. This study focuses solely on the additive outlier as it is the most common type found in time series due to its association with human error such as typing and recording mistakes. To understand the nature of an additive outlier, simulations of ARMA (1, 1) time series are contaminated with an additive outlier. Then, we examine the effect of the respected additive outlier on observations and residuals. It is found that an additive outlier does not affect observations prior to its existence and the ones subsequent to it. However, given the occurrence of the additive outlier at t=T, the additive outlier is seen to have a noticeable effect on the residual of t=T and t=T+1.

AB - Outliers are often encountered in time series analysis. Types of outliers that are usually dealt with are additive outlier, innovational outlier, temporary change and level shift. This study focuses solely on the additive outlier as it is the most common type found in time series due to its association with human error such as typing and recording mistakes. To understand the nature of an additive outlier, simulations of ARMA (1, 1) time series are contaminated with an additive outlier. Then, we examine the effect of the respected additive outlier on observations and residuals. It is found that an additive outlier does not affect observations prior to its existence and the ones subsequent to it. However, given the occurrence of the additive outlier at t=T, the additive outlier is seen to have a noticeable effect on the residual of t=T and t=T+1.

KW - ARMA (1

KW - Simulation Study

KW - The Nature of Additive Outlier

UR - http://www.scopus.com/inward/record.url?scp=67650631399&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67650631399&partnerID=8YFLogxK

M3 - Article

VL - 32

SP - 362

EP - 368

JO - European Journal of Scientific Research

JF - European Journal of Scientific Research

SN - 1450-202X

IS - 3

ER -