Results of this run
Affinity Propagation on Iris Flowers (damping 0.7) elects exemplar rows instead of taking k as input. Exemplar silhouette is 0.314, exemplar Calinski-Harabasz is 168.9, exemplar Davies-Bouldin is 0.976. Held-out species versus exemplars: adjusted Rand 0.298. 0 noise rows out of 150 without an exemplar. Exemplar group 0 on Iris Flowers holds 17 rows (petal length (cm) 1.40 / sepal length (cm) 4.69). Exemplar group 1 on Iris Flowers holds 11 rows (petal length (cm) 1.53 / sepal length (cm) 5.40). Exemplar group 2 on Iris Flowers holds 22 rows (petal length (cm) 1.48 / sepal length (cm) 5.05). Exemplar group 3 on Iris Flowers holds 24 rows (petal length (cm) 4.41 / sepal length (cm) 5.92). Exemplar group 4 on Iris Flowers holds 19 rows (petal length (cm) 3.94 / sepal length (cm) 5.53). Exemplar group 5 on Iris Flowers holds 14 rows (petal length (cm) 6.00 / sepal length (cm) 7.23). Exemplar group 6 on Iris Flowers holds 22 rows (petal length (cm) 5.17 / sepal length (cm) 6.50). Exemplar group 7 on Iris Flowers holds 3 rows (petal length (cm) 6.40 / sepal length (cm) 7.60). Exemplar group 8 on Iris Flowers holds 18 rows (petal length (cm) 5.17 / sepal length (cm) 6.23). Damping path: damping 0.5 elects 0.341 with 9 groups on Iris Flowers; damping 0.7 elects 0.314 with 9 groups on Iris Flowers; damping 0.9 elects 0.342 with 7 groups on Iris Flowers. Exemplar scatter keeps 95.8%.

| damping | Silhouette | Clusters |
|---|---|---|
| 0.5 | 0.341 | 9 |
| 0.7 | 0.314 | 9 |
| 0.9 | 0.342 | 7 |
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, Iris (Fisher, 1936). License: CC BY 4.0. The species 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.
