Dear Help,
I'm trying to calculate a weighted correlation matrix from a data frame with
6 columns (variables) and 297 observations extracted from the regression.
The last column is a weight column which I want to apply.
$ model :'data.frame': 297 obs. of 6 variables:
..$ VAR1 : num [1:297] 5.21 9.82 8.08 0.33 8.7 6.82 3.94 4 0 5 ...
..$ VAR2 : num [1:297] 7.82 9.79 8.5 5.59 7.82 7.13 5.59 7 0 5 ...
..$ VAR3 : num [1:297] 8.07 9.82 8.07 5.19 8.69 7.45 6.08 5 0 8 ...
..$ VAR4 : num [1:297] 6.66 9.81 8.62 4.97 8.62 5.9 5.9 0 0 6 ...
..$ VAR5 : num [1:297] 4.79 9.8 7.9 0.27 7.9 5.76 3.44 5 0 4 ...
..$ (weights): num [1:297] 1.01e-05 4.05e-01 1.24e-02 2.68e-05 3.02e-01
...
I use cov.wt() function (as this is the only function I could find to
calculate weighted correlations). If I use it without weight, it works OK:
> cov.wt(y$model[1:5],cor=TRUE)
$cov
VAR1 VAR2 VAR3 VAR4 VAR5
VAR1 7.250067 5.558081 5.342363 5.771774 6.049620
VAR2 5.558081 8.079245 6.616040 6.552380 5.122272
VAR3 5.342363 6.616040 7.985783 6.447208 4.867224
VAR4 5.771774 6.552380 6.447208 8.610053 5.077627
VAR5 6.049620 5.122272 4.867224 5.077627 6.717658
$center
VAR1 VAR2 VAR3 VAR4 VAR5
4.574175 4.450438 5.250236 4.439731 4.635421
$n.obs
[1] 297
$cor
VAR1 VAR2 VAR3 VAR4 VAR5
VAR1 1.0000000 0.7262204 0.7021075 0.7305256 0.8668584
VAR2 0.7262204 1.0000000 0.8236714 0.7856168 0.6952937
VAR3 0.7021075 0.8236714 1.0000000 0.7775173 0.6645285
VAR4 0.7305256 0.7856168 0.7775173 1.0000000 0.6676501
VAR5 0.8668584 0.6952937 0.6645285 0.6676501 1.0000000
If I try to apply weight, it doesn't work and gives me an error message:
> cov.wt(y$model[1:5],wt=y$model[6],cor=TRUE)
Error in cov.wt(y$model[1:5], wt = y$model[6], cor = TRUE) :
length of 'wt' must equal the number of rows in 'x'
The length of the weight variable is the same as for the other variables.
What do I do wrong?
Thanks in advance,
Lena
[[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.
