特征选择与稀疏学习

嵌入式选择与 L1 正则化

岭 ( Ridge ) 回归：

$$
\min_{w}\sum_{i = 1}^{m}\left ( y_{i} - w^{T}x_{i} \right )^{2}

• \lambda\left| \left| w \right| \right|_{2}^{2}\ \ ,\ \lambda > 0
  $$

LASSO 回归：

$$
\min_{w}\ \sum_{i = 1}^{m}\left ( y_{i} - w^{T}x_{i} \right )^{2} +
\lambda\left| \left| w \right| \right|_{1}\ ,\ \ \lambda > 0
$$