Table 2.

Indices used to evaluate the prediction method

Index	Definition and formula
Acc	(TP + TN)/(TP + TN + FP + FN)
AUC	area under the receiving operating characteristic curve
F-score	2 · TP/(2TP + FP + FN)
Sen	TP/(TP + FN)
Sp	TN/(TN + FP)
MCC	$$\frac{\mathit{TP}\cdot \mathit{TN}-\mathit{FN}\cdot \mathit{FP}}{\sqrt{\left(\mathit{TP}+\mathit{FN}\right)\cdot \left(\mathit{TP}+\mathit{FP}\right)\cdot \left(\mathit{TN}+\mathit{FN}\right)\cdot \left(\mathit{TP}+\mathit{FP}\right)}}$$

TP (true positive); TN (true negative); FP (false positive); FN (false negative);

Acc, overall accuracy; AUC, area under ROC; Sen, sensitivity; Sp, specificity; F-score, 2×precision×sensitivity/(precision+sensitivity); MCC, Matthews’s correlation coefficient.