glmnet: which is the reference category or class in multinomial regression?

Asked Nov 24 '14 at 10:30

Active Nov 24 '14 at 10:30

Viewed 1,431 times

Following post Why {glmnet} can be calculated parameters for all category?

I have 4 categories or classes or responses for y (thus multinomial): cat1, cat2, cat3 and finally no-cat. I want my "no category (no-cat)" class to be the reference category. Can I specify this in glmnet ? After performing a cv.glmnet I get coefficients for all categories.

Note: I'm a new glmnet user. Thanks in advance,

edited Apr 13 '17 at 12:44

Community

asked Nov 24 '14 at 10:30

guzu92

Related: http://stats.stackexchange.com/questions/121155/glmnet-how-to-make-sense-of-multinomial-parameterization – Andrew M Dec 03 '14 at 21:32
Thx, I'will check this post, seems like a goog glmnet test tutorial – guzu92 Dec 04 '14 at 14:18

glmnet: which is the reference category or class in multinomial regression?

0 Answers0