# what's the difference between multinomial logistic regression and traditional regression?

Could anyone please explain to me what is the difference between multinomial logistic regression and traditional regression?

I have used a method called elastic-net as the response variables are in discrete format, but I am confused if this method is suitable or not or should I be using a multinomial logistic regression instead?

I appreciated any help!

## 1 Answer

Multinomial logistic regression is used when the dependent variable is categorical with more than two levels but no order.

Linear regression is used when the dependent variable is continuous (or, at least, reasonably close to continuous).

Elastic net is a method of regularization. It can be used in both linear and logistic regression.

• +1 Succinct response to a broad question. For more information on each of the topics described above, I would recommend reading the sections on linear regression, multinomial logistic regression, and regularization in Introduction to Statistical Learning by James, Witten, Hastie, & Tibshirani www-bcf.usc.edu/~gareth/ISL/ISLR%20Seventh%20Printing.pdf – Mark White Aug 11 '18 at 15:11
• @MarkWhite Yeah, one could write a book on this (and some people have!) – Peter Flom Aug 11 '18 at 15:20
• @Peter Flom, I want to know how to be sure that the elastic net is suitable in both linear and logistic regression. I need a specific answer plz.appreciated your help! – F.caren Aug 12 '18 at 10:05
• You can check e.g. Wikipedia and the references therein. Or search for elastic net here on Cross Validated. – Peter Flom Aug 12 '18 at 10:20