[R] User defined covariate structure.

Spencer Graves spencer.graves at pdf.com
Sun Jul 30 16:07:29 CEST 2006


	  Have you tried using corARMA?  Won't this give you the symmetric 
Toeplitz form you desire, albeit in a different parameterization?

	  Hope this helps.
	  Spencer Graves

jswansmi at uoguelph.ca wrote:
> I am trying to use nlme but instead of using one of the “identity” variance or
> covariance matrixes such as compsymm or ar1.  Instead I want the covariance
> matrix to be represented in the following manor.  Is it possible to define my
> own covariance matrix?
> I have search and found papers saying I can define my own covariance matrixes
> and own correlation structures.  Said use corstruct but not sure how to
> implement it.  Also found documentation to use re.structur.  If able to help me
> out it be greatly appreciated as I am stuck.
> 
> |1	p1g	p2g	p3g	p4g	…|
> |p1g	1	p1g	p2g	p3g	…|
> |p2g 	p1g	1	p1g	p2g	…|
> |p3g	p2g	p1g	1	p1g 	…|
> |:	:	:	:	:	…|
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.



More information about the R-help mailing list