1) 精确度( precision ):TP / ( TP+FP ) = TP / P
2) 召回率(recall):TP / (TP + FN ) = TP / T
3) 真阳性率(True positive rate):TPR = TP / ( TP+FN ) = TP / T (敏感性 sensitivity)
4) 假阳性率(False positive rate):FPR = FP / ( FP + TN ) = FP / F (特异性:specificity)
5) 准确率(Accuracy):Acc = ( TP + TN ) / ( P +N )
6) F-measure:2*recall*precision / ( recall + precision )
7) ROC曲线:FPR为横坐标,TPR为纵坐标
8) PR曲线:recall为横坐标,precision 为纵坐标