[R] Model Averaging Help

Tom Finch tom.finch2 at btinternet.com
Thu May 3 15:51:52 CEST 2012


Dear All,

I'm using AIC-weighted model averaging to calculate model averaged 
parameter estimates and partial r-squares of each variable in a 
10-variable linear regression.

I've been using the MuMIn package to calculate parameter estimates, but 
am struggling with partial r-squares. There doesn't seem to be any 
function in the MuMIn package dealing with partial r-square 
(r.squaredLR() looks promising, but I think there is a problem updating 
MuMIn so I can't use this (new?) function). I know you can calculate 
partial r-squares using plot(anova(ols()), what = 'partial R2') but I 
don't know how best to do this for all 1024 models and then calculate 
the AIC-weighted model average.

Are there any packages out there that I'm missing? Or any ideas for 
combining MuMIn with anova.Design and Hmisc?

Thanks in advance, and apologies if my problem is badly worded.

Tom



More information about the R-help mailing list