Prediction of blood–brain barrier penetrating peptides based on data augmentation with Augur

Table 2 Performance of data augmentation results with different proportions

Evaluation strategy	Proportion	Sn	Sp	ACC	MCC	AUROC
Training set validation	0%	0.800	0.819	0.809	0.620	0.885
	25%	0.853	0.839	0.846	0.695	0.932
	50%	0.863	0.876	0.870	0.741	0.955
	75%	0.904	0.904	0.904	0.810	0.963
	100%	0.912	0.909	0.910	0.823	0.973
Independent set validation	0%	0.852	0.822	0.824	0.454	0.924
	25%	0.833	0.861	0.858	0.497	0.931
	50%	0.852	0.884	0.882	0.549	0.926
	75%	0.833	0.899	0.893	0.566	0.921
	100%	0.815	0.907	0.899	0.569	0.922

ISSN: 1741-7007