Simplified interpretation of accuracy and recall in AI papers

Reference resources ： Accuracy and recall @ A rookie makes up

Selected comments ： It's true , I really doubt that the person who defines this has a hole in his head , You said his first was Predictive value , The second is Truth value Is it not good? , Originally a very simple question , It's so complicated .

The accuracy rate is for our prediction results , It's about how many of the predicted positive samples are really positive samples . So there are two possibilities for a positive prediction , One is to predict a positive class as a positive class (TP), The other is to predict a negative class as a positive class (FP), That is to say

P=TPTP+FPP=\frac{TP}{TP+FP} P=TP+FPTP​

The recall rate is for our original sample , It shows how many positive examples in the sample are predicted correctly . There are two possibilities , One is to predict the original positive class into a positive class (TP), The other is to predict the original positive class as negative class (FN).

R=TPTP+FNR=\frac{TP}{TP+FN} R=TP+FNTP​

In fact, the denominator is different , A denominator is the number of samples predicted to be positive , The other is the number of all positive samples in the original sample .