Table 1 Comparative results of k-means, hierarchical, and DBSCAN clustering in the CVD prognostic datasets
From: Hyperparameter optimization for cardiovascular disease data-driven prognostic system
Clustering methods | K-means clustering | Hierarchical clustering | DBSCAN clustering |
|---|---|---|---|
The best number of clusters | 2 | 2 | 3 |
Distance metric | Euclidean | Euclidean | Euclidean |
Silhouette scores | C1 = -0.115 C2 = 0.097 | C1 = 0.128 C2 = -0.050 | C1 = 0.246 C2 = 0.000 C3 = 0.000 |