Table 9 Comparison of ADMM-based convex clustering-like method with DBSCAN

From: A parallel ADMM-based convex clustering method

Data set	No clust. ADMM	ADMM s.sc.	No clust. k-means	k-means s.sc.	DB- SCAN \(\epsilon\)	No clust. DB- SCAN	DB- SCAN s.sc.
\(30 \times 2\)	3	0.89	3	0.89	0.5	3	0.89
\(40 \times 2\)	6	0.39	5	0.57	0.8	3	0.65
\(1000 \times 3\)	8	0.76	8	0.76	2.5	8	0.75
\(5000 \times 3\)	4	0.77	4	0.78	2.5	4	0.76
\(5000 \times 5\)	5	0.62	4	0.75	5.0	4	0.75
\(10{,}000 \times 3\)	10	0.69	10	0.75	2.5	10	0.75

Back to article page