[R-sig-ME] "pdCompSymm" function in R package "nlme"

[Ben Bolker]

   Here are the available matrix classes available in nlme:

library(nlme)
apropos("^pd[^f]")
  [1] "pdBlocked"   "pdCompSymm"  "pdConstruct" "pdDiag"      "pdIdent"
  [6] "pdLogChol"   "pdMat"       "pdMatrix"    "pdNatural"   "pdSymm"

?pdClasses


  Unfortunately, I don't think there's anything useful in there:

pdIdent is a multiple of the identity matrix, pdDiag is (heterogeneous) 
diagonal.

I believe pdNatural, pdLogChol, pdSymm represent different 
parameterizations for the general positive-definite ("unstructured") 
correlation matrix.

   So "equal correlations but heterogeneous variances" doesn't appear to 
be an option.

   The MCMCglmm package has more options.
   glmmTMB offers a 'heterogeneous compound symmetric" but *not* (at 
present) a *homogeneous compound symmetric" option (although that could 
be achieved without too much difficulty by using the 'map' parameter to 
set some parameters equal ...)

On 10/6/21 11:51 AM, Isaac Kookhyun Yoo wrote:
> Dear all,
> 
> I am a graduate student. I have a question about pdCompSymm.
> pdCompSymm can be used for compound symmetry matrix only (All the variances
> are equal and all the covariances are equal.)
> 
> Are there any functions for unequal variances or unequal covariances matrix?
> 
> Thank you in advance.
> 
> Isaac
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-mixed-models using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
> 

-- 
Dr. Benjamin Bolker
Professor, Mathematics & Statistics and Biology, McMaster University
Director, School of Computational Science and Engineering
Graduate chair, Mathematics & Statistics