The scenario is as follows:
Suppose the original sample has two categories, True and False, where:
1. There are a total of T samples whose category is True;
2. There are a total of F samples whose category is False;
After classification prediction:
1. A total of TT samples whose category is True are judged as True by the system, and FT samples whose category is True are judged as False by the system, then TT+FT=T
2. A total of FF samples whose category is False are judged as False by the system, and TF samples whose category is False are judged as True by the system, then FF+TF=F
Indicator calculation:
Accuracy=TT/(TT+TF)--determine the proportion of true positive samples in positive samples
Accuracy rate=(TT+FF)/(T+F)--the proportion of correct judgment
Recall rate=TT/(TT+FT)--the proportion of correctly judged positive examples
False positive rate=FT/(TT+FT)--how many positive cases were missed
False alarm rate=TF/(TT+TF)-- reflects how many of the samples judged to be positive examples are negative examples
English sign:
Recall rate;
AccuracyPrecision;
Accuracy;
Missing Alarm;
False Alarm probability (False Alarm);