Page 1 of 1

Elastic Net Regression in EViews

Posted: Mon Sep 16, 2013 2:27 am
by ectrie
Hi,

Is there any way to implement the elastic net (a combination of the Lasso and Ridge regressions) in EViews, in which some variables are shrunken to 0 so that they are dropped from the model?
The estimation uses the objective function: min L(Beta) = the sum of( (y - X * Beta)^2 + lambda * P(Beta) ), in which the value of lambda is found using cross validation.

Thanks in advance!

Re: Elastic Net Regression in EViews

Posted: Mon Sep 16, 2013 7:29 am
by EViews Gareth
It is not built in.