APIS: accurate prediction of hot spots in protein interfaces by combining protrusion index with solvent accessibility

Table 2 Prediction performance of individual-feature based SVM models

Feature	Dataset	Specificity	Recall	Precision	Accuracy	F1	TP	TN	FP	FN
RcsASA	Training set	0.79	0.74	0.71	0.77	0.72	46	73	19	16
	Test set	0.66	0.67	0.46	0.66	0.55	26	58	30	13
RctASA	Training set	0.78	0.71	0.69	0.75	0.70	44	72	20	18
	Test set	0.68	0.72	0.50	0.69	0.59	28	60	28	11
RcpASA	Training set	0.78	0.79	0.71	0.79	0.75	49	72	20	13
	Test set	0.70	0.59	0.47	0.67	0.52	23	62	26	16
BsRASA	Training set	0.72	0.79	0.65	0.75	0.72	49	66	26	13
	Test set	0.52	0.72	0.40	0.58	0.51	28	46	42	11
RcsmPI	Training set	0.75	0.81	0.68	0.77	0.74	50	69	23	12
	Test set	0.74	0.69	0.54	0.72	0.61	27	65	23	12
BtRASA	Training set	0.72	0.69	0.62	0.71	0.66	43	66	26	19
	Test set	0.56	0.72	0.42	0.61	0.53	28	49	39	11
BpRASA	Training set	0.62	0.82	0.59	0.70	0.69	51	57	35	11
	Test set	0.53	0.67	0.39	0.57	0.49	26	47	41	13
RctmPI	Training set	0.76	0.73	0.67	0.75	0.70	45	70	22	17
	Test set	0.78	0.67	0.58	0.75	0.62	26	69	19	13
BsASA	Training set	0.61	0.81	0.58	0.69	0.68	50	56	36	12
	Test set	0.61	0.59	0.40	0.61	0.48	23	54	34	16

ISSN: 1471-2105