Dear R-List,
I would like to have a large number of stratified random subsamples drawn from
my dataframe and automatically test for correlation differences in every
subsample.
Let this be my dataframe
df<-data.frame(group=c(rep(1,5),rep(2,5),rep(3,5)),a=c(3,4,5,6,3,4,5,4,5,4,1,2,1,2,1),b=c(1,2,3,4,5,3,4,3,4,5,6,5,6,2,3),c=c(2,2,3,3,5,1,1,6,6,5,6,1,1,2,1))
Then I would like to have n subsamples with one row out of each group, e.g.
> df
group a b c
1 1 3 1 2
2 1 4 2 2
3 1 5 3 3
4 1 6 4 3
5 1 3 5 5
6 2 4 3 1
7 2 5 4 1
8 2 4 3 6
9 2 5 4 6
10 2 4 5 5
11 3 1 6 6
12 3 2 5 1
13 3 1 6 1
14 3 2 2 2
15 3 1 3 1
>df.sub1
group a b c
1 3 1 2
2 5 4 1
3 1 3 1
>df.sub2
group a b c
1 4 2 2
2 4 3 6
3 2 2 2
etc.
And then test, if the correlation ab is significantly higher than the
correlation ac.
I managed to perform the test "manually" using r.test from the "psych" package,
however I did not succeed in doing it automatically, i.e. I had
to do cor(df.sub) for all subsamples an put the values manually into the
r.test-code (which is very time consuming if you have to do it 100 times). Is
there a nice way to combine the stratified subsampling with a code that can do
the r.test with dataframe input directly (I mean without me entering all
correlations ab, ac, bc manually)?
Thank you for any hint!
Alain
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org 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.
