Closed philip-fu closed 6 years ago
Generally, there is no way to do this. It's possible to implement some differentiation approximation of AUC and optimize over it (e.g. pairwise loss), but for now, it is not implemented.
Generally, there is no way to do this. It's possible to implement some differentiation approximation of AUC and optimize over it (e.g. pairwise loss), but for now, it is not implemented.