Computational models for the prediction of adverse cardiovascular drug reactions

Table 5 Provides the overall performance measures for the models generated using biological, chemical, and phenotypic features and the combination of the two and three levels of features on non-redundant testing dataset using over sampling of minority class

Type of feature	RF						SMO
Type of feature	ACC	Precision	Recall	F-score	AUC	PRC	ACC	Precision	Recall	F-score	AUC	PRC
Biological	93.56	0.94	0.99	0.96	0.52	0.93	91.24	0.93	0.99	0.96	0.51	0.93
Chemical	91.41	0.94	0.97	0.95	0.52	0.94	88.75	0.94	0.95	0.93	0.48	0.93
Phenotypic	93.83	0.95	1.00	0.97	0.50	0.93	93.83	0.94	1.00	0.97	0.50	0.93
Biological + chemical	90.24	0.94	0.96	0.94	0.53	0.94	89.06	0.94	0.95	0.94	0.53	0.93
Biological + phenotypic	93.66	0.94	1.00	0.97	0.52	0.94	93.30	0.94	0.99	0.96	0.52	0.93
Chemical + phenotypic	91.49	0.94	0.97	0.95	0.50	0.94	90.54	0.94	0.96	0.95	0.48	0.93
Biological + chemical + phenotypic	90.92	0.94	0.96	0.95	0.54	0.95	89.07	0.94	0.95	0.94	0.47	0.93

ISSN: 1479-5876