Biometrika Advance Access originally published online on August 12, 2008
Biometrika 2008 95(3):773-778; doi:10.1093/biomet/asn023
Miscellanea |
A note on conditional AIC for linear mixed-effects models
Department of Biostatistics and Computational Biology, University of Rochester Medical Center, Rochester, New York 14642, U.S.A. hliang{at}bst.rochester.edu hwu{at}bst.rochester.edu
Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100080, China guohua_zou{at}urmc.rochester.edu
Received for publication 1 September 2006. Revision received 1 January 2008.
The conventional model selection criterion, the Akaike information criterion, AIC, has been applied to choose candidate models in mixed-effects models by the consideration of marginal likelihood. Vaida & Blanchard (2005) demonstrated that such a marginal AIC and its small sample correction are inappropriate when the research focus is on clusters. Correspondingly, these authors suggested the use of conditional AIC. Their conditional AIC is derived under the assumption that the variance-covariance matrix or scaled variance-covariance matrix of random effects is known. This note provides a general conditional AIC but without these strong assumptions. Simulation studies show that the proposed method is promising.
Key Words: Akaike information criterion • Conditional likelihood • Longitudinal data • Marginal likelihood • Mixed-effects model • Model selection
References
-
Akaike H. Information theory and an extension of the maximum likelihood principle. In: Second International Symposium on Information Theory—Petrov B., Csaki F., eds. (1973) Budapest: Akademiai Kiado. 267–81.
Burnham K. P., Anderson D. P. Model Selection and Inference: A Practical Information-Theoretical Approach (1998) New York: Springer.
Hodges J. S., Sargent D. J. Counting degrees of freedom in hierarchical and other richly parameterized models. Biometrika (2001) 88:367–79.
Hurvich C. M., Simonoff J. S., Tsai C. L. Smoothing parameter selection in nonparametric regression using an improved Akaike information criterion. J. R. Statist. Soc. (1998) 60:271–93.[CrossRef]
Hurvich C. M., Tsai C. L. Regression and time series model selection in small samples. Biometrika (1989) 76:297–307.
Laird N. M., Ware J. H. Random effects models for longitudinal data. Biometrika (1982) 38:963–74.
Lu K. L., Berger J. O. Estimation of normal means: Frequentist estimation of loss. Ann. Statist. (1989) 17:890–906.[CrossRef]
Naik P. A., Tsai C. L. Single-index model selections. Biometrika (2001) 88:821–32.
Ngo L., Brand R. Model selection in linear mixed effects models using SAS Proc Mixed. (2002) SAS Global Forum 22.
Pinheiro J. C., Bates D. M. Mixed-Effects Models in S and S-PLUS (2000) New York: Springer.
Schwarz G. Estimating the dimension of a model. Ann. Statist. (1978) 6:461–4.[CrossRef]
Stein C. M. Estimation of the mean of a multivariate normal distribution. Ann. Statist. (1981) 9:1135–51.[CrossRef]
Vaida F., Blanchard S. Conditional Akaike information for mixed-effects models. Biometrika (2005) 92:351–70.
Verbeke G., Molenberghs G. Linear Mixed Models for Longitudinal Data (2000) New York: Springer.
Vonesh E. F., Chinchilli V. M. Linear and Nonlinear Models for the Analysis of Repeated Measurements (1996) New York: Marcel Dekker.
Ye J. M. On measuring and correcting the effects of data mining and model selection. J. Am. Statist. Assoc. (1998) 93:120–31.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||