Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Cameron McIntosh <cnm100@hotmail.com> |
To | STATA LIST <statalist@hsphsun2.harvard.edu> |
Subject | RE: st: model selection using information criteria with xtlsdvc or xtabond2 |
Date | Tue, 22 Nov 2011 21:59:02 -0500 |
Sebastian, Perhaps you can use the natural log of the residual sum of squares, e.g., AIC = n*ln(SSR) + 2*k BIC = n*ln(SSR) + k*ln(n) where n is the sample size and k is the number of estimated parameters. Yamaoka, K., Nakagawa, T., & Uno, T. (1978). Application of Akaike’s information criterion (AIC) in the evaluation of linear pharmacokinetic equations. Journal of Pharmacokinetics and Biopharmaceutics, 6, 165–175. Bonate, P.L. (2011). Pharmacokinetic-Pharmacodynamic Modeling and Simulation (2nd ed.). New York, NY: Springer. My two cents, Cam ---------------------------------------- > From: sebastian.petrick@ifw-kiel.de > Date: Tue, 22 Nov 2011 17:34:24 -0800 > Subject: st: model selection using information criteria with xtlsdvc or xtabond2 > To: statalist@hsphsun2.harvard.edu > > Dear statalisters, > > I am running a highly unbalanced large-T, small-N (T: ~30, N: ~50-150) > dynamic panel estimation using Bruno's xtlsdvc estimator. Naturally, > the estimator doesn't provide information on the log-likelihood (as > does xtabond/xtabond2). I still would like to do model selection using > the standard information criteria, like Akaike's AIC or the BIC. Is > there a well-functioning work-around avoiding the use of the > log-likelihood? > > Thanks > > Sebastian > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/