Greetings,
I have recently been exploring the 'glmnet' package and subsequently
cv.glmnet. The basic code as follows:
model <- cv.glmnet(variables, group, family="multinomial", alpha=.5,
standardize=F)
I understand that cv.glmnet does k-fold cross-validation to return a value
of lambda. However, sometimes when I follow up the cv.glmnet to extract
the coefficients either very few or all are zero. If I understand this
correctly, it means that there aren't very many (if any) variables to
separate the groups. Despite this, I would like to provide a list of
variables and rank them in terms of importance (even if not discriminatory
as this is for some simulation purposes and not working on a particular
question/experiment). Is there a way for my to set up the analysis to
provide a user determined number of variables? Or perhaps another way, is
it possible to determine the order with which variables are dropped from
the model?
Best regards,
--
Charles Determan
Integrated Biosciences PhD Candidate
University of Minnesota
[[alternative HTML version deleted]]
______________________________________________
[email protected] 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.
