SVM 合页损失函数 导数
Lets use the example of the SVM loss function for a single datapoint:We can differentiate the function with respect to the weights. For example, taking the gradient with respect to Wyi we obtain