From: Dr Carbon <drcarbon_at_gmail.com>

Date: Sun 06 Mar 2005 - 01:08:41 EST

language R

R-help@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 Received on Sun Mar 06 01:20:04 2005

Date: Sun 06 Mar 2005 - 01:08:41 EST

I'm trying to get the coefficient of partial determination for each of three independent variables. I've tried mvr in package pls.pcr. I'm a little confused by the output. I'm curious how I can order the LV's according to their names rather than their relative contribution to the regression.

For instance, using the crabs data from MASS I made a regression of FL~RW+noise

set.seed(124)

library(pls.pcr)

library(MASS)

attach(crabs)

crabs.simpls <- mvr(data.frame(x1 = RW,x2 = runif(200)), FL,
validation="CV", method="SIMPLS")

summary(crabs.simpls)

crabs.simpls <- mvr(data.frame(x1 = runif(200), x2 = RW), FL,
validation="CV", method="SIMPLS")

summary(crabs.simpls)

# compare to summary(lm(FL~RW+runif(200)))
detach(crabs)

The two summaries are almost identical, as are the inputs. But the order of the LVs are different. How can I know that it is x1 is the useful predictor in the first example and that x2 is the useful predictor in the second example. I hope to run a three variable regression in a MC framework and output the partial rsq for x1, x2, and x3 in every run.

Can I do this? I fear I've made some fundamental misunderstanding about mvr()

Thanks, DC

> version

arch i386 os mingw32 system i386, mingw32 status major 2 minor 0.1 year 2004 month 11 day 15

language R

R-help@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 Received on Sun Mar 06 01:20:04 2005

*
This archive was generated by hypermail 2.1.8
: Fri 03 Mar 2006 - 03:30:41 EST
*