Meta-learning-based Inductive logistic matrix completion for prediction of kinase inhibitors

Table 7 Performance comparison of various methods on tail kinase of LTKinaseDB (local evaluation model)

	AUC	AUPR	BA	RECALL	PRECISION	F1
MetaILMC (MACCS + CTD)	0.8754	0.7265	0.8215	0.9170	0.6432	0.6636
MetaILMC (ECFP + ProtVec)	0.8461	0.6726	0.7972	0.8896	0.5429	0.6018
ILMC	0.7724	0.5401	0.7319	0.8196	0.4726	0.5202
MTDNN [20]	0.7403	0.5044	0.6971	0.8605	0.4383	0.4906
SVM	0.5541	0.5153	0.5541	0.4599	0.3925	0.3016
KNN	0.5586	0.4958	0.5586	0.6097	0.2916	0.3554
RF	0.5976	0.5139	0.5976	0.6090	0.3263	0.3870
MolTrans [29]	0.6366	0.6245	0.6124	0.8042	0.4862	0.4220
MetaMGNN[30]	0.7280	0.6294	0.7280	0.8037	0.4334	0.5099

ISSN: 1758-2946