Computational models for the prediction of adverse cardiovascular drug reactions

Table 7 Provides the overall performance measures for the models generated using eighty biological, chemical, and phenotypic features and the combination of the two and three levels of features on non-redundant testing dataset

Type of feature	RF						SMO
Type of feature	ACC	Precision	Recall	F-score	AUC	PRC	ACC	Precision	Recall	F-score	AUC	PRC
Biological	93.76	0.94	0.99	0.96	0.48	0.94	93.59	0.93	0.99	0.96	0.50	0.93
Chemical	93.03	0.93	0.98	0.95	0.54	0.95	93.98	0.93	100.00	0.96	0.50	0.93
Phenotypic	94.85	0.95	0.98	0.97	0.69	0.96	94.23	0.94	0.99	0.96	0.52	0.94
Biological + chemical	93.81	0.94	0.98	0.96	0.55	0.95	93.94	0.93	0.99	0.96	0.50	0.93
Biological + phenotypic	94.31	0.94	0.99	0.96	0.72	0.96	94.57	0.94	0.99	0.97	0.50	0.94
Chemical + phenotypic	94.97	0.95	0.99	0.97	0.72	0.97	94.16	0.94	0.99	0.95	0.52	0.94
Biological + chemical + phenotypic	94.28	0.94	0.99	0.96	0.68	0.96	92.34	0.95	0.99	0.97	0.51	0.95

ISSN: 1479-5876