| |
True
|
Class
|
| |
P
|
N
|
Hypothesized |
Y
|
True Positive
|
False Positive
|
Class
|
N
|
False Nagative
|
Trure Nagative
|
P = TP + FN 正样本
N = TN + FP 负样本
TPR = TP / P
FPR = FP / N
Precision = TP / (TP + FP)
Recall = TP / P
Accuary = (TP + TN) / (P + N)
F-measure = 2Pre * Rec / (Pre + Rec)