Table 2.

Summary of statistical parameters for the k-NN classification models

Entry	Fingerprints	NER	k	Sensitivity		Specificity
				Class 1	Class 2	Class 1	Class 2
1	CDK
	Fitting	0.68	1	0.72	0.65	0.65	0.72
	CV	0.66	1	0.72	0.61	0.61	0.72
	External	0.54	1	0.52	0.57	0.57	0.52
2	Estate
	Fitting	0.68	1	0.73	0.62	0.62	0.73
	CV	0.66	1	0.72	0.61	0.61	0.72
	External	0.53	1	0.49	0.57	0.57	0.49
3	Extended CDK
	Fitting	0.67	1	0.70	0.63	0.63	0.70
	CV	0.65	1	0.70	0.61	0.61	0.70
	External	0.56	1	0.56	0.57	0.57	0.56
4	CDK graph
	Fitting	0.64	1	0.69	0.59	0.59	0.69
	CV	0.64	1	0.70	0.58	0.58	0.70
	External	0.55	1	0.52	0.57	0.57	0.52
5	MACCS
	Fitting	0.68	6	0.76	0.59	0.59	0.76
	CV	0.67	6	0.76	0.57	0.57	0.76
	External	0.55	6	0.54	0.55	0.55	0.54
6	PubChem
	Fitting	0.60	3	0.69	0.52	0.52	0.69
	CV	0.60	3	0.71	0.49	0.49	0.71
	External	0.57	3	0.62	0.52	0.52	0.62
7	Sub-structure
	Fitting	0.68	1	0.70	0.67	0.67	0.70
	CV	0.67	1	0.69	0.66	0.66	0.69
	External	0.57	1	0.54	0.59	0.59	0.54
8	Sub-structure count
	Fitting	0.67	1	0.74	0.61	0.61	0.74
	CV	0.68	1	0.72	0.65	0.65	0.72
	External	0.58	1	0.61	0.56	0.56	0.61