✦ LIBER ✦

A Note on Maximum Likelihood Ratio Test of No Outliers in Regression Models

✍ Scribed by Prof. S. R. Paul

Publisher: John Wiley and Sons
Year: 1986
Tongue: English
Weight: 180 KB
Volume: 28
Category: Article
ISSN: 0323-3847
DOI: 10.1002/bimj.4710280809

No coin nor oath required. For personal study only.

✦ Synopsis

A derivation of the maximum likelihood ratio test for testing no outliere in regreeeion models h given ueing the method of WETEXEILL (1981, pp. 106-107) for estimating the regreeeion parsmetere. This method h eseentially eimilar to the one outlined in B a s m and Lmwm (1978, p. 283), although by our detailed deriv8tion it ie easier to me that the maximum likelihood estimate of f! of model (3) under the hypotheeis that the ith obwrvation in an outlier in the name aa that obtained from model (1) when the ith obeervation.is removed.

📜 SIMILAR VOLUMES

Finite Sample Properties of the Maximum

Finite Sample Properties of the Maximum Likelihood Estimator and of Likelihood Ratio Tests in Hidden Markov Models

✍ S. Michalek; M. Wagner; J. Timmer; W. Vach 📂 Article 📅 2001 🏛 John Wiley and Sons 🌐 English ⚖ 170 KB 👁 2 views

Estimation of a Conditional Mean in a Li

Estimation of a Conditional Mean in a Linear Regression Model after a Preliminary Test on Regression Coefficient

✍ Dinesh S. Bhoj; Mohammad Ahsanullah 📂 Article 📅 1994 🏛 John Wiley and Sons 🌐 English ⚖ 449 KB 👁 1 views

## Abstract Consider the two linear regression models of __Y__~__ij__~ on __X__~__ij__~, namely __Y__~__ij__~ = β~__io__~ + β~__ij__~, __X__~__ij__~ + __E__~__ij__~ = 1, 2,…, __n__~__i__~, __i__ = 1, 2, where __E__~__ij__~ are assumed to be normally distributed with zero mean and common unknown var