ElasticNet

• ==Combines both of lasso and ridge==
• It will try to learn a sparse model like Lasso and retain regularization properties of ridge
• Tuned with $\alpha$ and $l1_{ratio}$ parameters