Table 3.

Model statistics for CYP2C9 isoform. Models are described according to the method and type of descriptors, the Applicability Domain (AD: yes/no), number of variables (p) and classification parameters (parameter: object/leaf ratio for CART, k for kNN and α for N3). For each model, the Non-Error Rate (NER), the Sensitivity (Sn), and the Specificity (Sp) are reported in Fitting, Cross-Validation and on the test set. %out indicates the percentage of test set compounds outside the AD.

Model	Descriptors	AD	p	Parameter	Fitting			Cross-Validation			Test Set
					NER	Sn	Sp	NER	Sn	Sp	%out	NER	Sn	Sp
CART	MD	y	4	210	0.75	0.75	0.75	0.75	0.75	0.75	-	0.75	0.75	0.74
		n	4	210	0.75	0.75	0.75	0.75	0.75	0.75	0	0.75	0.75	0.74
k-NN	MD	y	6	14	0.77	0.69	0.85	0.77	0.68	0.85	-	0.76	0.67	0.86
		n	6	14	0.77	0.69	0.85	0.77	0.68	0.85	5	0.76	0.67	0.84
N3	ECFP	y	1024	1	0.80	0.87	0.73	0.80	0.86	0.73	-	0.78	0.83	0.73
		n	1024	1	0.80	0.87	0.73	0.80	0.86	0.73	1	0.78	0.83	0.73