Table 1.

Statistical Significance Analysis of Model Performance using ROC Scores.

Machine Learning Methods	p-Value vs. KA–PI MTDNN	Average ROC
KA–PI LR	<2.2 × 10−16	0.887
KA–PI RF	<2.2 × 10−16	0.842
KA–PI NB	<2.2 × 10−16	0.852
KA–PI MTDNN		0.958

Abbreviations: LR = logistic regression; RF = random forest; NB = naïve Bayes; MTDNN = multi-task deep neural networks.