Results of this run
Affinity Propagation on Wine (UCI wine recognition) (damping 0.7) elects exemplar rows instead of taking k as input. Exemplar silhouette is 0.120, exemplar Calinski-Harabasz is 24.3, exemplar Davies-Bouldin is 1.638. Held-out wine_class versus exemplars: adjusted Rand 0.245. 0 noise rows out of 178 without an exemplar. Exemplar group 0 on Wine (UCI wine recognition) holds 17 rows (proline 1220.7 / magnesium 97.06). Exemplar group 1 on Wine (UCI wine recognition) holds 4 rows (proline 753.8 / magnesium 125.0). Exemplar group 2 on Wine (UCI wine recognition) holds 19 rows (proline 925.2 / magnesium 101.2). Exemplar group 3 on Wine (UCI wine recognition) holds 10 rows (proline 1290.5 / magnesium 114.7). Exemplar group 4 on Wine (UCI wine recognition) holds 15 rows (proline 1045.3 / magnesium 112.8). Exemplar group 5 on Wine (UCI wine recognition) holds 6 rows (proline 624.2 / magnesium 96.33). Exemplar group 6 on Wine (UCI wine recognition) holds 3 rows (proline 801.7 / magnesium 149.7). Exemplar group 7 on Wine (UCI wine recognition) holds 16 rows (proline 501.7 / magnesium 87.50). Exemplar group 8 on Wine (UCI wine recognition) holds 17 rows (proline 478.4 / magnesium 88.71). Exemplar group 9 on Wine (UCI wine recognition) holds 3 rows (proline 440.7 / magnesium 91.67). Exemplar group 10 on Wine (UCI wine recognition) holds 17 rows (proline 447.8 / magnesium 90.06). Exemplar group 11 on Wine (UCI wine recognition) holds 21 rows (proline 653.1 / magnesium 96.57). Exemplar group 12 on Wine (UCI wine recognition) holds 6 rows (proline 535.0 / magnesium 118.3). Exemplar group 13 on Wine (UCI wine recognition) holds 9 rows (proline 605.6 / magnesium 97.00). Exemplar group 14 on Wine (UCI wine recognition) holds 15 rows (proline 626.5 / magnesium 99.40). Damping path: damping 0.5 elects 0.121 with 14 groups on Wine (UCI wine recognition); damping 0.7 elects 0.120 with 15 groups on Wine (UCI wine recognition); damping 0.9 elects 0.120 with 15 groups on Wine (UCI wine recognition). Exemplar scatter keeps 55.4%.

| damping | Silhouette | Clusters |
|---|---|---|
| 0.5 | 0.121 | 14 |
| 0.7 | 0.120 | 15 |
| 0.9 | 0.120 | 15 |
from sklearn.cluster import AffinityPropagation
from sklearn.preprocessing import StandardScaler
Xs = StandardScaler().fit_transform(X)
labels = AffinityPropagation(damping=0.7, random_state=42).fit_predict(Xs)
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.
