Sensitivity and Specificity Improvement in Abdominal Obesity Diagnosis Using Cluster Analysis during Waist Circumference Cut-Off Point Selection

Table 4

Distribution of the subjects (training and testing datasets) during cross-validation process.

	Cluster Analysis in S2
	Cluster 1		Cluster 2		Cluster 3		Total
		%		%		%		%
Cluster Analysis in S2 according to S1
Cluster 1	609	64.0	0	0	31	3.3	640	67.3
Cluster 2	0	0	50	5.3	0	0	50	5.3
Cluster 3	0	0	14	1.5	247	26.0	261	27.4
Total	609	64.0	64	6.7	278	29.2	951	100.0

S1: a subgroup of subjects randomly selected from the database.
S2: a subgroup of subjects randomly selected from the database.
Cohen’s kappa coefficient: 0.902 ( < 0.00001).