Encyclopedia entry
Partial Likelihood
Wiley StatsRef: Statistics Reference Online, pp.1-5
John Wiley & Sons, Ltd
12/11/2018
DOI: 10.1002/9781118445112.stat05932
Abstract
While the traditional likelihood paradigm has been central to much of statistical inference, modified versions of the likelihood function have been developed to address situations where the full likelihood might be difficult to formulate or optimize. One such variant arises in dealing with models characterized by high-dimensional parameter sets. In these frameworks, it is sometimes possible to appropriately partition the likelihood so that a factor can be isolated that depends on a parameter subset of key inferential interest. This factor, which is only a portion of the full likelihood, defines the partial likelihood. Many of the standard asymptotic properties of maximum likelihood estimators also hold for estimators that maximize the partial likelihood. We discuss the general development of the partial likelihood and illustrate its utility in modeling frameworks involving both survival and time series data.
Details
- Title: Subtitle
- Partial Likelihood
- Creators
- Javier E Flores - University of Iowa, BiostatisticsJoseph E Cavanaugh - University of Iowa Iowa City IA USA
- Contributors
- N Balakrishnan (Editor)Theodore Colton (Editor)Brian Everitt (Editor)Walter Piegorsch (Editor)Fabrizio Ruggeri (Editor)Jozef L Teugels (Editor)
- Resource Type
- Encyclopedia entry
- Publication Details
- Wiley StatsRef: Statistics Reference Online, pp.1-5
- DOI
- 10.1002/9781118445112.stat05932
- Publisher
- John Wiley & Sons, Ltd; Chichester, UK
- Number of pages
- 2
- Comment
- Update based on original article by David Harrington, Wiley StatsRef: Statistics Reference Online, ©2014, John Wiley & Sons, Ltd.
- Language
- English
- Date published
- 12/11/2018
- Academic Unit
- Statistics and Actuarial Science; Biostatistics; Injury Prevention Research Center
- Record Identifier
- 9984214671202771
Metrics
177 Record Views