Please use this identifier to cite or link to this item: https://scholar.utcc.ac.th/handle/6626976254/3278
Title: Confidence Intervals for Autoregressive Parameter in AR(1) with Outliers by Improved Recursive Median OLS Estimator
Authors: Saothayanun, Luckhana 
Kanjanasakda, Yupin 
Somrang, Boonying 
Taweesakulvatchara, Sunee 
Issue Date: 2010
Publisher: Chulalongkorn University Printing House
University of the Thai Chamber of Commerce
Source: Luckhana Saothayanun, Yupin Kanjanasakda, Boonying Somrang, Sunee Taweesakulvatchara (2010) Confidence Intervals for Autoregressive Parameter in AR(1) with Outliers by Improved Recursive Median OLS Estimator. University of the Thai Chamber of Commerce Journal Vol.30 No.4.
Journal: University of the Thai Chamber of Commerce Journal 
Abstract: This research aims to compare the efficiency of the confidence interval for theautoregressive parameter (ρ) in a first-order autoregressive model with outliers. Theconfidence interval for the autoregressive parameter (ρ) in a first-order autoregressivemodel with outliers is estimated by the Recursive Mean Adjustment Estimator (RMA).The estimator is proposed by replacing mean with median, namely the RecursiveMedian Adjustment Estimator (RMD), and the new estimator becomes the ImprovedRecursive Median Adjustment Estimator (IRMD). In this study, only the additive outlierswere investigated. The percentages of outliers are 1% and 5%. The magnitude ofoutliers is equal to 3 and 5 times the standard deviation. The sample sizes are 25, 50,100 and 250. The results of the research for small and moderate sample sizes (n = 25 and 50) and five percentages of outliers are as follows. The confidence intervalfor the autoregressive parameter (ρ) by IRMD estimator assigns more value tocoverage probability than do RMA and RMD estimators in almost all ρ values. Forlarge sample sizes (n = 100 and 250), the confidence interval for the autoregressiveparameter (ρ) by IRMD estimator assigns more value to coverage probability thando RMA and RMD estimators in almost all ρ values, except when the percentageof outliers is 1% and the magnitude of outliers is equal to 3σ.
URI: https://scholar.utcc.ac.th/handle/6626976254/3278
ISSN: 0125-2437
Rights: This work is protected by copyright. Reproduction or distribution of the work in any format is prohibited without written permission of the copyright owner.
Appears in Collections:JEO: Journal Articles

Files in This Item:
File Description SizeFormat 
546fulltext.pdf689.11 kBAdobe PDFThumbnail
View/Open
Show full item record Recommend this item

Page view(s)

14
checked on Jul 11, 2019

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.