评估指标：ROC，AUC，Precision、Recall、F1-score

一、ROC，AUC

ROC（Receiver Operating Characteristic）曲线和AUC常被用来评价一个二值分类器（binary classifier）的优劣 。

ROC曲线一般的横轴是FPR，纵轴是FPR。AUC为曲线下面的面积，作为评估指标，AUC值越大，说明模型越好。如下图：

 

二、Precision、Recall、F1-score

Terminology and derivations
from a  confusion matrix

true positive (TP) eqv. with hit true negative (TN) eqv. with correct rejection false positive (FP) eqv. with  false alarm,  Type I error false negative (FN) eqv. with miss,  Type II error sensitivity or  true positive rate (TPR) eqv. with  hit rate,  recall specificity (SPC) or  true negative rate (TNR) precision or  positive predictive value (PPV) negative predictive value (NPV) fall-out or  false positive rate (FPR) false discovery rate (fdr) miss rate or  false negative rate (FNR) accuracy (ACC) F1 score is the  harmonic mean of  precision and  sensitivity Matthews correlation coefficient (MCC) Informedness = Sensitivity + Specificity - 1 Markedness = Precision + NPV - 1

中文解释