View : 99 Download: 0

A modified LGB method for detecting significant effects based on a half-normal probability plot

Title
A modified LGB method for detecting significant effects based on a half-normal probability plot
Authors
Chung, Jong HeeLim, Yong Bin
Ewha Authors
임용빈
SCOPUS Author ID
임용빈scopus
Issue Date
2019
Journal Title
JOURNAL OF THE KOREAN STATISTICAL SOCIETY
ISSN
1226-3192JCR Link

1876-4231JCR Link
Citation
JOURNAL OF THE KOREAN STATISTICAL SOCIETY vol. 48, no. 4, pp. 568 - 577
Keywords
Detecting significant effectsHalf-normal probability plotAdjusted box plot
Publisher
KOREAN STATISTICAL SOC
Indexed
SCIE; SCOPUS; KCI WOS scopus
Document Type
Article
Abstract
In analyzing data from unreplicated factorial designs, the half-normal probability plot is commonly used to screen for the 'vital few' effects. Recently, many formal methods have been proposed to overcome the subjectivity of this plot. Lawson (1998) (hereafter denoted as LGB) suggested a hybrid method based on the half-normal probability plot, which is a blend of Lenth (1989) and Loh (1992) method. The method consists of fitting a simple least squares line to the inliers, which are determined by the Lenth method. The effects exceeding the prediction limits based on the fitted line are candidates for the vital few effects. To improve the accuracy of partitioning the effects into inliers and outliers, we propose a modified LGB method (hereafter denoted as the Mod_LGB method), in which more outliers can be classified by using both the Carling's modification of the box plot (Carling, 2000) and Lenth method. If no outlier exists or there is a wide range in the inliers as determined by the Lenth method, more outliers can be found by the Carling method. A simulation study is conducted in unreplicated 2(4) designs with the number of active effects ranging from 1 to 6 to compare the efficiency of the Lenth method, original LGB methods, and the proposed modified version of the LGB method. (C) 2019 The Korean Statistical Society. Published by Elsevier B.V. All rights reserved.
DOI
10.1016/j.jkss.2019.01.002
Appears in Collections:
자연과학대학 > 통계학전공 > Journal papers
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

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

BROWSE