In ridge regression, however, the formula for the hat matrix should include the regularization penalty: Hridge = X(X′X + λI) −1 X, which gives dfridge = trHridge ...
確定! 回上一頁