✦ LIBER ✦

On the existence of maximum likelihood estimates in random effects models for clustered multivariate binary data

✍ Scribed by David Todem; KyungMann Kim

Publisher: Springer
Year: 2009
Tongue: English
Weight: 237 KB
Volume: 73
Category: Article
ISSN: 0026-1335
DOI: 10.1007/s00184-009-0263-2

No coin nor oath required. For personal study only.

📜 SIMILAR VOLUMES

On the existence and uniqueness of maxim

On the existence and uniqueness of maximum likelihood estimates in polytomous response models

✍ Siu-Keung Tse 📂 Article 📅 1986 🏛 Elsevier Science 🌐 English ⚖ 282 KB

A note on the existence of maximum likel

A note on the existence of maximum likelihood estimates for Gaussian-inverted Wishart models

✍ N. Le; L. Sun; J.V. Zidek 📂 Article 📅 1998 🏛 Elsevier Science 🌐 English ⚖ 221 KB

We show that a non-identifiability problem can occur when one attempts to estimate the degrees of freedom and the unstructured covariance matrix simultaneously in a Gaussian-inverted Wishart model, using the maximum likelihood approach. However, the EM algorithm may falsely converge to give finite e

Approximate solutions for the maximum-li

Approximate solutions for the maximum-likelihood estimates in models of univariate human twin data

✍ U. W. L. Wijesiri; Christopher J. Williams 📂 Article 📅 1995 🏛 Springer US 🌐 English ⚖ 407 KB

A multivariate approach for estimating t

A multivariate approach for estimating the random effects variance component in one-way random effects model

✍ Brajendra C. Sutradhar 📂 Article 📅 1997 🏛 Elsevier Science 🌐 English ⚖ 416 KB

The effect of kurtosis in the estimation

The effect of kurtosis in the estimation of the parameters of the one-way random effects model from familial data

✍ M.M. Shoukri; D.S. Tracy; I.U.H. Mian 📂 Article 📅 1990 🏛 Elsevier Science 🌐 English ⚖ 414 KB

On the propriety of a modified Jeffreys'

On the propriety of a modified Jeffreys's prior for variance components in binary random effects models

✍ Ranjini Natarajan 📂 Article 📅 2001 🏛 Elsevier Science 🌐 English ⚖ 96 KB

This paper proves that a modiÿed Je reys's prior on the variance components in binary random e ects models is integrable under mild conditions on the link function. These conditions are shown to be satisÿed for two commonly used link functions, the logit and probit functions.