TruePositive+FalsePostive & F-Measure

1. TruePositive + FalsePostive + FalseNegative + TrueNegative


The illustration of four metrics in Pattern Recognition are as follows:

1)True Positive: A sample is a positve one and is classified as Positive;

2)False Negative: A sample is a positive one and is classified as Negative;

3)True Negative: A sample is a negative one and is classified as Negative;

4)False Negative: A sample is a negative one and is classified as Positive.

                                                    Predicted Class Label              

                                                P                                       N                    

                                          ___________________________________   

Actual Class Label   P       True  Positive(TP)            False Negative(FN)

                                          ___________________________________    

                                N       False Positive(FP)           True   Negative(TN)


2. Recall + Precision


Recall: related to the samples with actual label of Positive; it is calculated with the formula:

              Recall(%) = TP/(TP+FN)x100%

Precision:related to the samples classified as Positive; it is calculated with the formula:

              Precision(%) = TP/(TP+FP)x100% 


3. F-Measure

F-Measure is another metric for describe the classification performance of the system. It is calculated through the Recall (R) and Precision (P) rate as follows:


          F = {(a*a + 1)P*R}/{a*a(P+R)}

When a = 1, the measure becomes F1-Measure:

          F1 = 2P*R/(P+R)




阅读更多
个人分类: 计算机视觉
想对作者说点什么? 我来说一句

没有更多推荐了,返回首页

加入CSDN,享受更精准的内容推荐,与500万程序员共同成长!
关闭
关闭