Evaluation Criteria——评价标准

PRE=TP/(TP+FP)           precision

REC=TP/(TP+FN)          recall

ACU=(TP+TN)/(TP+FP+TN+FN)  accuracy

F=(2*PRE*REC)/(PRE+REC)     F-score

 

True positive: Positives correctly identified as positives     实际为阳性样本中被正确判断为阳性的比例
False positive: Negatives incorrectly identified as positives   实际为阴性样本中被错误判断为阳性的比例
True negative: Negatives correctly identified as negatives    实际为阴性样本中被正确判断为阴性的比例
False negative: Positives incorrectly identified as negatives  实际为阳性样本中被错误判断为阴性的比例

posted on 2017-02-19 20:43  修行君  阅读(972)  评论(0)    收藏  举报

导航