Table 2.

Performance scores at FPR = 0.20 for 21 prediction models by 5-fold CV with the training dataset that was consisted of 1:2 ratio of positive and negative samples.

Predictors classifier (encoding)	TPR	TNR	FNR	ACC	MCC	MCR	AUC	pAUC
ADA (CKSAAP)	0.676 (0.32)	0.800 (0.00)	0.323 (0.32)	0.689 (0.01)	0.378 (0.16)	0.311 (0.01)	0.737 (0.04)	0.12 (0.06)
ADA (binary)	0.613 (0.03)	0.800 (0.01)	0.386 (0.03)	0.657 (0.01)	0.315 (0.03)	0.343 (0.01)	0.718 (0.03)	0.11 (0.07)
ADA (AAC)	0.644 (0.31)	0.801 (0.00)	0.355 (0.31)	0.692 (0.03)	0.383 (0.02)	0.291 (0.09)	0.747 (0.05)	0.133 (0.04)
ADA (CKSAAP, binary)	0.650 (0.24)	0.800 (0.01)	0.349 (0.24)	0.702 (0.12)	0.407 (0.24)	0.297 (0.12)	0.771 (0.10)	0.136 (0.05)
ADA (CKSAAP, AAC)	0.661 (0.09)	0.800 (0.00)	0.339 (0.09)	0.712 (0.10)	0.417 (0.21)	0.289 (0.10)	0.783 (0.09)	0.139 (0.03)
ADA (binary, AAC)	0.653 (0.12)	0.800 (0.00)	0.347 (0.12)	0.710 (0.13)	0.412 (0.11)	0.292 (0.13)	0.778 (0.10)	0.137 (0.09)
ADA (CKSAAP, binary, AAC)	0.654 (0.08)	0.801 (0.00)	0.346 (0.08)	0.721 (0.01)	0.456 (0.15)	0.287 (0.07)	0.799 (0.02)	0.140 (0.06)
SVM (CKSAAP)	0.677 (0.16)	0.800 (0.00)	0.323 (0.03)	0.712 (0.12)	0.425 (0.07)	0.287 (0.02)	0.788 (0.06)	0.143 (0.07)
SVM (binary)	0.683 (0.02)	0.800 (0.00)	0.317 (0.03)	0.718 (0.01)	0.438 (0.15)	0.281 (0.09)	0.787 (0.08)	0.138 (0.04)
SVM (AAC)	0.681 (0.12)	0.801 (0.00)	0.316 (0.12)	0.704 (0.13)	0.382 (0.06)	0.325 (0.01)	0.785 (0.03)	0.134 (0.05)
SVM (CKSAAP, binary)	0.711 (0.08)	0.800 (0.00)	0.293 (0.29)	0.728 (0.11)	0.445 (0.26)	0.256 (0.11)	0.799 (0.23)	0.146 (0.09)
SVM (CKSAAP, AAC)	0.543 (0.13)	0.802 (0.00)	0.456 (0.09)	0.667 (0.23)	0.376 (0.12)	0.356 (0.04)	0.800 (0.03)	0.154 (0.11)
SVM (binary, AAC)	0.567 (0.12)	0.801 (0.00)	0.432 (0.11)	0.684 (0.13)	0.382 (0.21)	0.324 (0.12)	0.803 (0.05)	0.169 (0.10)
SVM (CKSAAP, binary, AAC)	0.598 (0.08)	0.802 (0.00)	0.401 (0.13)	0.700 (0.12)	0.422 (0.09)	0.312 (0.02)	0.812 (0.07)	0.170 (0.08)
RF (CKSAAP)	0.798 (0.15)	0.800 (0.00)	0.201 (0.15)	0.749 (0.26)	0.500 (0.20)	0.251 (0.11)	0.803 (0.16)	0.145 (0.08)
RF (binary)	0.735 (0.09)	0.800 (0.00)	0.264 (0.09)	0.721 (0.14)	0.443 (0.01)	0.278 (0.02)	0.793 (0.13)	0.143 (0.06)
RF (AAC)	0.691 (0.15)	0.801 (0.00)	0.308 (0.15)	0.786 (0.26)	0.584 (0.20)	0.213 (0.11)	0.791 (0.16)	0.141 (0.08)
RF (CKSAAP, binary)	0.806 (0.02)	0.800 (0.00)	0.193 (0.01)	0.754 (0.02)	0.510 (0.06)	0.246 (0.10)	0.823 (0.03)	0.158 (0.02)
RF (CKSAAP, AAC)	0.681 (0.13)	0.800 (0.00)	0.319 (0.13)	0.659 (0.23)	0.502 (0.18)	0.182 (0.09)	0.797 (0.14)	0.151 (0.08)
RF (binary, AAC)	0.725 (0.09)	0.802 (0.00)	0.275 (0.09)	0.671 (0.14)	0.588 (0.01)	0.185 (0.02)	0.826 (0.13)	0.159 (0.06)
RF (CKSAAP, binary, AAC)	0.810 (0.02)	0.802 (0.00)	0.190 (0.01)	0.778 (0.02)	0.666 (0.06)	0.141 (0.10)	0.832 (0.03)	0.168 (0.02)

Better results with each of ADA, SVM and RF were highlighted by bold values.

The values within the first bracket indicate the standard error (SE).