We transform the original m-class classification problem to (m+1)-class where the (m+1)-th class represents the model abstaining from making ...
確定! 回上一頁