损失函数

Posted 2020-08-09 iAthena

1 Square loss

    

 

2 Hinge loss

 

3 Logistic loss

 

4 Cross entropy loss

   Using the alternative label convention,so that,

    the cross entropy loss is defined as

      

 

x = np.linspace(-3, 3, 1000)
y_logit = np.log(1 + np.exp(-x)) / math.log(2)
y_01 = x < 0
y_hinge = 1 - x
y_hinge[y_hinge < 0] = 0

 

来自为知笔记(Wiz)