DeepAR: a novel deep learning-based hybrid framework for the interpretable prediction of androgen receptor antagonists

Table 3 Independent test results of different stacked models on the independent test dataset

Meta-model	ACC	Sn	Sp	MCC	AUC	F1
DeepAR	0.911	0.897	0.927	0.823	0.945	0.912
RF	0.888	0.862	0.915	0.777	0.941	0.888
LGBM	0.888	0.885	0.890	0.775	0.947	0.890
SVMRBF	0.864	0.851	0.878	0.728	0.913	0.865
XGB	0.864	0.839	0.890	0.729	0.949	0.864
LR	0.893	0.874	0.915	0.788	0.949	0.894
PLS	0.893	0.874	0.915	0.788	0.951	0.894
NB	0.888	0.862	0.915	0.777	0.906	0.888
ET	0.899	0.897	0.902	0.799	0.952	0.902
MLP	0.876	0.839	0.915	0.754	0.927	0.874
KNN	0.834	0.816	0.854	0.670	0.835	0.835
SVMLN	0.846	0.816	0.878	0.694	0.913	0.845
ADA	0.858	0.828	0.890	0.718	0.922	0.857
DT	0.822	0.816	0.829	0.645	0.823	0.826

ISSN: 1758-2946