Hi R users
I have a 1000 * 11 matrix (or data.frame). The first 10 columns consist of 0
or 1 (i.e., binary data) and the last column consists of real values.
I would like to run a logistic regression with the first binary column as a
dependent variable, and the second binary column and the last real value
column as independent variables. After this, I save 'Z' value for the
binary
independent variable. Then, with the first binary column as a dependent
variable, and the third binary column and the last real value column as
independent variables. After this, I saved 'Z' value for the binary
independent variable, keeping doing this ...
Here is my simple R codes
logistic.z.value <-c()
for (i in 1:9)
{
for (j in (i+1):10)
{
model <- glm(X[,i] ~ X[,11] + X[,j], family=binomial(logit))
logistic.z.value <-
append(logistic.z.value,coef(summary(model))[3,3])
}
}
I would like to avoid using 'for' loop command. Is there any way for me
to
use some fast codes instead of using 'for' loop?
Many thanks
Taka,
_________________________________________________________________
Stay in touch with old friends and meet new ones with Windows Live Spaces