--- Renaud Lancelot <[EMAIL PROTECTED]> wrote: > Hi Peng, > > Peng wrote: > [snip] > > I am still wondering whether there is a function > like > > getVarCov(), so that I can get R var-cov matrix > > directly. I looked through the function list of > nlme, > > but I cannot find. > > > > Peng > > R>library(nlme) > R>data(Orthodont) > R>fm1 <- lme(distance ~ age, data = Orthodont) # > random is ~ age > R> > R>lapply(pdMatrix(fm1$modelStruct$reStruct), "*", > fm1$sigma^2) > $Subject > (Intercept) age > (Intercept) 5.414722 -0.32102403 > age -0.321024 0.05126664 > > Comment: pdMatrix returns the list of var-cov > matrices of the random > effects (one matrix for each level of the possibly > multilevel > structure). For some computational reason, these > matrices are scaled by > the residuals variance: you have to multiply them by > the residuals variance. > > If you really want a matrix (the above is a list > with one component), > you can do: > > R>lapply(pdMatrix(fm1$modelStruct$reStruct), "*", > fm1$sigma^2)$Subject > (Intercept) age > (Intercept) 5.414722 -0.32102403 > age -0.321024 0.05126664
Hi, Renaud, I think reStruct gives the structure of random effects, not the within group errors. And I happened to want the structure for the later one, especially when the within group errors are differently structured between groups. Regards, Peng ------------------------------ Peng Liu | Division of Statistics | Northern Illinois University | De Kalb, IL 60115, USA | E-mail: [EMAIL PROTECTED] | ------------------------------ > > > Best, > > Renaud > > -- > Dr Renaud Lancelot, v�t�rinaire > CIRAD, D�partement Elevage et M�decine V�t�rinaire > (CIRAD-Emvt) > Programme Productions Animales > http://www.cirad.fr/fr/pg_recherche/page.php?id=14 > > ISRA-LNERV tel +221 832 49 > 02 > BP 2057 Dakar-Hann fax +221 821 18 > 79 (CIRAD) > Senegal e-mail > [EMAIL PROTECTED] > ______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
