[R] correlation structure

From: Daniel O'Shea <dan.oshea_at_dnr.state.mn.us>
Date: Mon, 25 Jun 2007 15:05:44 -0500


I have been using a nlme model and wish to specify the correlation structure. My data is grouped (bas), but I have no time component or adequate spatial description beyond the grouping variable. So I chose the simplest structure and updated my original nlme model by: update(model, corr=corCompSymm(~1|bas))

This is not significant. One of my variables measures area upstream of the sample. Area is a continuous variable, and there are some identical values of area. Can I model the dependence among residuals with in group (bas) and along area? Can the corCAR1 be used and the area used as a position variable? If so does the position variable need to be sorted when using the corCAR1 structure? Or could one use spatial classes and use area as one position and watershed as the other substituting numbers for the grouping variable? I have been trying, but R keeps crashing when I specify correlation structures other than compound symmetry. I assume this is because I am not specifying the form correctly. Any suggestions would be appreciated.

Dan



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. Received on Mon 25 Jun 2007 - 20:12:41 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Mon 25 Jun 2007 - 21:32:42 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.