Skip to main content
. 2021 Apr 12:bbab111. doi: 10.1093/bib/bbab111

Table 1.

Performance of different types of ML algorithms and DL for prediction of PPI modulators

ML Algorithm Positives Negatives CV Sens Spec FPR Prec F1-Score MCC ROC AUC PRC AUC
RF 993 993 2-Fold 0.77 0.82 0.18 0.81 0.79 0.60 0.87 0.87
993 993 10-Fold 0.84 0.88 0.12 0.88 0.86 0.72 0.93 0.92
993 993 LOO 0.84 0.88 0.12 0.88 0.86 0.72 0.93 0.92
NaiveBayes 993 993 2-Fold 0.69 0.77 0.23 0.73 0.73 0.46 0.81 0.80
993 993 10-Fold 0.70 0.81 0.19 0.75 0.75 0.50 0.83 0.82
993 993 LOO 0.69 0.80 0.20 0.75 0.75 0.50 0.83 0.82
SMO 993 993 2-Fold 0.73 0.68 0.32 0.70 0.71 0.42 0.71 0.65
993 993 10-Fold 0.76 0.72 0.28 0.74 0.74 0.47 0.74 0.74
993 993 LOO 0.76 0.72 0.28 0.74 0.74 0.49 0.74 0.68
SMO-RBF 993 993 2-Fold 0.71 0.83 0.17 0.77 0.77 0.54 0.77 0.71
993 993 10-Fold 0.77 0.84 0.16 0.83 0.81 0.62 0.81 0.75
993 993 LOO 0.78 0.85 0.15 0.81 0.81 0.62 0.81 0.75
Deeplearning 993 993 2-Fold 0.75 0.68 0.32 0.71 0.72 0.43 0.78 0.76
993 993 10-Fold 0.77 0.72 0.28 0.74 0.74 0.48 0.81 0.79
993 993 LOO 0.75 0.72 0.28 0.73 0.73 0.47 0.80 0.79