. 2021 Mar 10;159:107217. doi: 10.1016/j.csda.2021.107217

Table 3.

Percentage of numbers of clusters identified correctly (IC) based on 1000 simulation replications when data are generated (27) with respect to $k$ -means (K), convex clustering (Convex), and $k$ -means++ (KPP) directly on regression coefficients based on the gap statistic and our BIC selector in generalized $k$ -means.

$σ$	$c$	$k$	$n_{0} = 50$				$n_{0} = 100$
			K	Convex	KPP	BIC	K	Convex	KPP	BIC
0.1	10	2	92.6	98.5	88.2	$100.0$	96.5	$100.0$	93.9	$100.0$
		3	72.3	99.2	90.2	$100.0$	72.4	99.9	93.1	$100.0$
	20	2	99.7	$100.0$	98.9	$100.0$	$100.0$	$100.0$	99.9	$100.0$
		3	73.3	$100.0$	99.2	$100.0$	72.9	$100.0$	99.9	$100.0$
0.2	10	2	94.6	97.8	90.9	$100.0$	96.0	99.8	93.3	$100.0$
		3	36.8	47.5	26.9	$100.0$	75.6	99.2	95.0	$100.0$
	20	2	99.7	$100.0$	99.0	$100.0$	$100.0$	$100.0$	99.9	$100.0$
		3	24.6	24.4	10.6	$100.0$	77.2	$100.0$	99.8	$100.0$
0.5	10	2	95.7	99.9	93.5	$100.0$	96.0	99.1	95.4	$100.0$
		3	1.1	1.3	1.2	$88.7$	1.3	1.4	1.4	$100.0$
	20	2	99.9	$100.0$	99.7	$100.0$	99.9	$100.0$	99.8	$100.0$
		3	0.0	0.0	0.0	$95.2$	0.0	0.0	0.0	$100.0$
1.0	10	2	96.2	$100.0$	93.0	$100.0$	97.7	$100.0$	95.8	$100.0$
		3	0.9	0.8	1.6	$16.2$	0.3	0.2	0.3	$57.5$
	20	2	99.6	$100.0$	99.6	$100.0$	$100.0$	$100.0$	99.9	$100.0$
		3	0.0	0.0	0.0	$39.8$	0.0	0.0	0.0	$84.3$