Skip to main content
. 2017 Apr 18;7:897. doi: 10.1038/s41598-017-00952-9

Figure 6.

Figure 6

(a) ROC curves showing the relationship between TPR (sensitivity) and FPR (1-specificity) for the three ML algorithms used for model building. Most optimal results were obtained by RF with an accuracy of 84% and the AUC of 0.81. (b) 10-fold cross-validation performance measures of three classifiers.