Table 2.

Performance comparison of our deep learning architecture and three sequence encoding schemes-based RF methods using 5-fold cross-validation

	AUPRC
Human–virus PPI dataset	Our method	LD+RF	CT+RF	AC+RF
Human–HIV	0.974	0.972	0.970	0.972
Human–Herpes	0.768	0.741	0.737	0.699
Human–Papilloma	0.818	0.740	0.724	0.656
Human–Influenza	0.834	0.813	0.795	0.713
Human–Hepatitis	0.636	0.571	0.580	0.537
Human–Dengue	0.605	0.526	0.505	0.456
Human–Zika	0.746	0.720	0.718	0.698
Human–SARS-CoV-2	0.329	0.371	0.350	0.314