Can Regularization regressions be implemented in Eviews?
Posted: Tue May 17, 2011 3:58 pm
I am interesting for regressions where there are constraints regarding the betas.
They are thought to be known as regularization regressions and the most common are Lasso and Ridge Regressions
This means that they try to evaluate the minimum error of residuals as the regular regression but the beta are constrained.
For example in Ridge regression there is a constraint of: Σ(β^2)<=S , S>0
In lasso regressions there is a constraint of : Σ(|βi|)<=S, S>0
Thank you in advance!
They are thought to be known as regularization regressions and the most common are Lasso and Ridge Regressions
This means that they try to evaluate the minimum error of residuals as the regular regression but the beta are constrained.
For example in Ridge regression there is a constraint of: Σ(β^2)<=S , S>0
In lasso regressions there is a constraint of : Σ(|βi|)<=S, S>0
Thank you in advance!