Table 4.

Comparison of precision and recall (%) for three classification models—Naïve Bayes, linear discriminant and Adaboost. Results correspond to patients that were not included in training set for leave-one-patient-out testing.

Patient	Naïve Bayes		Linear discriminant		Adaboost
	Precision	Recall	Precision	Recall	Precision	Recall
P1	0	0	0	0	0	0
P2	81	50	0	0	0	0
P3	93	79	96	86	98	64
P4	98	94	98	53	0	0
P5	44	100	58	100	94	100
P6	93	100	97	100	0	0