Identification of infectious disease-associated host genes using machine learning techniques

BMC Bioinformatics

Table 4 Performance on imbalanced datasets using deep neural network classifier

Features set	Vector length	P(+): N(−)	Sensitivity (%)	Specificity (%)	Accuracy (%)	PPV (%)	MCC	F1 score (%)	AUC
Selected Features For PAAC _Network properties	16	1: 1	85.61	86.57	86.33	86.91	0.73	86.15	0.899
Selected Features For PAAC _Network properties	16	1: 2	77.89	92.56	87.81	84.64	0.72	80.72	0.900
Selected Features For PAAC _Network properties	16	1: 3	72.34	94.54	89.03	81.70	0.70	76.53	0.902
Selected Features For PAAC _Network properties	16	1: 4	68.89	95.46	90.20	79.20	0.68	73.52	0.897
Selected Features For PAAC _Network properties	16	1: 5	69.00	95.13	90.85	74.44	0.66	71.25	0.895
Selected Features For Normalized And Filtered PAAC_ Network properties	10	1: 1	84.62	87.63	86.44	88.06	0.73	86.00	0.894
Selected Features For Normalized And Filtered PAAC_ Network properties	10	1: 2	76.76	92.94	87.62	84.41	0.72	80.25	0.895
Selected Features For Normalized And FilteredPAAC_ Network properties	10	1: 3	74.35	93.52	88.91	80.40	0.70	76.88	0.895
Selected Features For Normalized And Filtered PAAC_ Network properties	10	1: 4	67.39	96.27	90.57	82.68	0.69	73.66	0.897
Selected Features For Normalized And Filtered PAAC_ Network properties	10	1: 5	67.52	96.01	91.31	77.95	0.67	71.97	0.895

ISSN: 1471-2105