Skip to main content

Table 4 Values obtained for clustering metrics Silhouette Score. Calinski-Harabasz Index and Davies-Bouldin Score for Hierarchical Clustering with Gower’s Distance using k=[4]. For Silhouette Score and Calinski-Harabasz a higher value indicates a better performance, for Davies-Bouldin a lower value is best

From: Identifying subgroups in heart failure patients with multimorbidity by clustering and network analysis

Clusters

Silhouette Score

Calinski-Harabasz

Davies-Bouldin

2

0.273

1147.505

1.494

3

0.153

918.620

1.859

4

0.155

910.162

1.751

5

0.153

810.238

1.797

6

0.147

746.957

1.739

7

0.141

673.593

1.983

8

0.127

615.439

2.005

9

0.125

571.963

2.041

10

0.130

538.409

1.939

11

0.132

512.288

1.961

12

0.140

491.961

1.914