A note on NPML estimation for exponential family regression models with unspecified dispersion parameter

Einbeck, Jochen; Hinde, John

Abstract

Nonparametric maximum likelihood (NPML) estimation for exponential families with unspecified dispersion parameter \phi suffers from computational instability, which can lead to highly fluctuating EM trajectories and suboptimal solutions, in particular when \phi is allowed to vary over mixture components. In this paper, a damped version of the EM algorithm is proposed to cope with these problems.

Citation

Einbeck, J., & Hinde, J. (2006). A note on NPML estimation for exponential family regression models with unspecified dispersion parameter. Austrian Journal of Statistics, 35(2&3), 233-243

• Journal Article Type: Article
• Publication Date: Jun 1, 2006
• Deposit Date: Sep 29, 2008
• Publicly Available Date: Sep 29, 2008
• Journal: Austrian Journal of Statistics
• Print ISSN: 1026-597X
• Publisher: Austrian Society for Statistics
• Peer Reviewed: Peer Reviewed
• Volume: 35
• Issue: 2&3
• Pages: 233-243
• Keywords: EM algorithm, Random effect models, Nonparametric maximum likelihood, Overdispersion, Gamma distribution, Generalized linear model.
• Public URL: https://durham-repository.worktribe.com/output/1547822
• Publisher URL: http://www.stat.tugraz.at/AJS/ausg062+3/Welcome.html