Predicting potential drug-drug interactions by integrating chemical, biological, phenotypic and network data

BMC Bioinformatics

Table 3 Performances of different models evaluated by 20 runs of 3-CV

Method	Similarity	Index	AUC	AUPR	Recall	Precision	Accuracy	F
Neighbor recommender Method	Substructure	1	0.935	0.808	0.772	0.669	0.927	0.717
	Target	2	0.806	0.425	0.420	0.579	0.831	0.486
	Transporter	3	0.714	0.405	0.344	0.495	0.800	0.406
	Enzyme	4	0.753	0.437	0.466	0.424	0.853	0.443
	Pathway	5	0.810	0.624	0.674	0.510	0.898	0.581
	Indication	6	0.903	0.640	0.584	0.658	0.888	0.618
	Label	7	0.935	0.803	0.758	0.673	0.925	0.713
	Off label	8	0.939	0.815	0.771	0.684	0.928	0.725
	CN	9	0.940	0.816	0.761	0.691	0.927	0.724
	AA	10	0.941	0.816	0.761	0.690	0.927	0.724
	RA	11	0.942	0.819	0.763	0.691	0.928	0.725
	Katz	12	0.933	0.782	0.715	0.666	0.917	0.689
	ACT	13	0.866	0.721	0.629	0.574	0.915	0.600
	RWR	14	0.940	0.814	0.760	0.688	0.927	0.722
Random walk Method	Substructure	15	0.935	0.807	0.768	0.670	0.927	0.716
	Target	16	0.844	0.608	0.601	0.555	0.888	0.576
	Transporter	17	0.713	0.437	0.339	0.504	0.795	0.404
	Enzyme	18	0.760	0.533	0.655	0.374	0.886	0.476
	Pathway	19	0.810	0.648	0.724	0.515	0.906	0.601
	Indication	20	0.939	0.820	0.773	0.693	0.930	0.731
	Label	21	0.936	0.809	0.771	0.674	0.927	0.719
	Off label	22	0.937	0.811	0.771	0.680	0.928	0.722
	CN	23	0.937	0.807	0.748	0.685	0.925	0.715
	AA	24	0.937	0.806	0.747	0.683	0.924	0.714
	RA	25	0.936	0.799	0.741	0.675	0.923	0.706
	Katz	26	0.936	0.801	0.743	0.677	0.923	0.708
	ACT	27	0.866	0.706	0.658	0.699	0.834	0.643
	RWR	28	0.938	0.813	0.759	0.690	0.927	0.723
Matrix perturbation method		29	0.941	0.813	0.755	0.709	0.928	0.731

ISSN: 1471-2105