Table 2.

Results of comparing the clustering approaches

Clustering Approach	No. of hospitals			CH-index*	Silhouette**	Davies-Bouldin***
Size	Small: 853	Medium: 201	Large: 70	647.35	0.48	1.08
Ownership	Non-profit: 450	Private: 238	Public: 436	25.77	-0.11	7.59
SOM	Cluster 1: 186	Cluster 2: 249	Cluster 3: 689	874.54	0.57	0.76

^* A high score is achieved when clusters are dense and well separated.

^** The score ranges from $-1$ for incorrect clustering to $+1$ for dense and well-separated clustering.

^*** A value closer to zero indicates a better partition.