Results of this run
PAM medoids on Wine (UCI wine recognition) pick 3 real Wine (UCI wine recognition) rows as centers (k=3). Medoid silhouette is 0.268, medoid Calinski-Harabasz is 67.1, medoid Davies-Bouldin is 1.425, medoid inertia is 500.93. Held-out wine_class matches the medoids at adjusted Rand 0.741. 0 noise rows out of 178 were still assigned. Medoid group 0 on Wine (UCI wine recognition) holds 55 rows (proline 488.7 / magnesium 90.60). Medoid group 1 on Wine (UCI wine recognition) holds 49 rows (proline 627.6 / magnesium 99.02). Medoid group 2 on Wine (UCI wine recognition) holds 74 rows (proline 1017.9 / magnesium 107.0). Medoid sweep: k 2 reaches 0.268 with 2 groups on Wine (UCI wine recognition); k 3 reaches 0.268 with 3 groups on Wine (UCI wine recognition); k 4 reaches 0.187 with 4 groups on Wine (UCI wine recognition); k 5 reaches 0.154 with 5 groups on Wine (UCI wine recognition); k 6 reaches 0.174 with 6 groups on Wine (UCI wine recognition); k 7 reaches 0.106 with 7 groups on Wine (UCI wine recognition); k 8 reaches 0.112 with 8 groups on Wine (UCI wine recognition). The medoid scatter is PCA at 55.4% variance.

| k | Silhouette | Clusters |
|---|---|---|
| 2 | 0.268 | 2 |
| 3 | 0.268 | 3 |
| 4 | 0.187 | 4 |
| 5 | 0.154 | 5 |
| 6 | 0.174 | 6 |
| 7 | 0.106 | 7 |
| 8 | 0.112 | 8 |
from sklearn.preprocessing import StandardScaler
Xs = StandardScaler().fit_transform(X)
# PAM with k-medoids++ init, euclidean distance, random_state=42, n_clusters=3.
# The sandbox K-Medoids control uses the same seed and scaling.
Who, how, and why
Method: StandardScaler, drop constant columns. Code version 3e4afc4b4352 (scripts/seo/build_explore_artifacts.py). scikit-learn 1.4.2. Seed 42. Dataset source: UCI Machine Learning Repository, wine recognition. License: CC BY 4.0. The wine_class column was held out of the fit and used only for the adjusted Rand score.
Results computed on 2026-11-02.
The results paragraph is filled from these numbers by a fixed template in the site code, so the scores can be checked against the same dataset in the tool.
Maintainer: nikolaou1982.
Contact: nick1982100@gmail.com.
