In the previous section, we saw Ridge regression: this is a method for regularization and for avoiding overfitting. In Ridge regression, the regression coefficients are shrunk by introducing a penalty, as follows:
Here, the term λβ12 is a shrinkage penalty that decreases when the β parameters withdraw (shrink) towards zero. The Lasso regression is a shrinkage method like Ridge, with subtle but important differences. The Lasso estimate is defined by the following equation:
Here, the term λ|β1| is a shrinkage penalty for the ...