A procedure for testing suspected observations

The problem of multiple outliers detection in one-parameter exponential family is considered. The outlier detection procedure involves two estimates of scale parameter which are obtained by maximizing two log-likelihoods; the complete data log-likelihood and its conditional expectation given suspected observations. The procedure is also applied to the exponential and normal samples. Copyright Springer-Verlag 2013

MoreLess

Year of publication:	2013
Authors:	Kumar, Nirpeksh
Published in:	Statistical Papers. - Springer. - Vol. 54.2013, 2, p. 471-478
Publisher:	Springer
Subject:	Outlier \| Discordancy \| One-parameter exponential family \| Log-likelihood

More details

Extent:	text/html
Type of publication:	Article
Source:	RePEc - Research Papers in Economics

Persistent link: https://www.econbiz.de/10010680672