Conditional Akaike information under generalized linear and proportional hazards mixed models
- 13 July 2011
- journal article
- research article
- Published by Oxford University Press (OUP) in Biometrika
- Vol. 98 (3) , 685-700
Abstract
We study model selection for clustered data, when the focus is on cluster specific inference. Such data are often modelled using random effects, and conditional Akaike information was proposed in Vaida & Blanchard (2005) and used to derive an information criterion under linear mixed models. Here we extend the approach to generalized linear and proportional hazards mixed models. Outside the normal linear mixed models, exact calculations are not available and we resort to asymptotic approximations. In the presence of nuisance parameters, a profile conditional Akaike information is proposed. Bootstrap methods are considered for their potential advantage in finite samples. Simulations show that the performance of the bootstrap and the analytic criteria are comparable, with bootstrap demonstrating some advantages for larger cluster sizes. The proposed criteria are applied to two cancer datasets to select models when the cluster-specific inference is of interest.Keywords
This publication has 32 references indexed in Scilit:
- Partitioning Degrees of Freedom in Hierarchical and Other Richly Parameterized ModelsTechnometrics, 2010
- A note on conditional AIC for linear mixed-effects modelsBiometrika, 2008
- Model selection for multi‐component frailty modelsStatistics in Medicine, 2007
- Measuring the complexity of generalized linear hierarchical modelsCanadian Journal of Statistics, 2007
- Extended information criterion (EIC) approach for linear mixed effects models under restricted maximum likelihood (REML) estimationStatistics in Medicine, 2005
- Random effects Cox models: A Poisson modelling approachBiometrika, 2003
- Bayesian Measures of Model Complexity and FitJournal of the Royal Statistical Society Series B: Statistical Methodology, 2002
- Counting degrees of freedom in hierarchical and other richly-parameterised modelsBiometrika, 2001
- Bootstrapping Likelihood for Model Selection with Small SamplesJournal of Computational and Graphical Statistics, 1999
- Tests for Variation Over Groups in Survival DataJournal of the American Statistical Association, 1995