Results of this run
BIRCH CF-tree on Wine (UCI wine recognition) (threshold 0.5, then k=3, threshold=0.5) scores silhouette 0.277, CF-tree Calinski-Harabasz 67.6, CF-tree Davies-Bouldin 1.419. Held-out wine_class versus the leaf clustering: adjusted Rand 0.790. 0 noise rows out of 178 in the tree. CF leaf 0 on Wine (UCI wine recognition) holds 58 rows (proline 501.4 / magnesium 92.55). CF leaf 1 on Wine (UCI wine recognition) holds 56 rows (proline 624.9 / magnesium 99.86). CF leaf 2 on Wine (UCI wine recognition) holds 64 rows (proline 1076.0 / magnesium 106.2). BIRCH k path: k 2 splits to 0.267 with 2 groups on Wine (UCI wine recognition); k 3 splits to 0.277 with 3 groups on Wine (UCI wine recognition); k 4 splits to 0.226 with 4 groups on Wine (UCI wine recognition); k 5 splits to 0.187 with 5 groups on Wine (UCI wine recognition); k 6 splits to 0.180 with 6 groups on Wine (UCI wine recognition); k 7 splits to 0.187 with 7 groups on Wine (UCI wine recognition); k 8 splits to 0.188 with 8 groups on Wine (UCI wine recognition). CF-tree scatter keeps 55.4%.

| k | Silhouette | Clusters |
|---|---|---|
| 2 | 0.267 | 2 |
| 3 | 0.277 | 3 |
| 4 | 0.226 | 4 |
| 5 | 0.187 | 5 |
| 6 | 0.180 | 6 |
| 7 | 0.187 | 7 |
| 8 | 0.188 | 8 |
from sklearn.cluster import Birch
from sklearn.preprocessing import StandardScaler
Xs = StandardScaler().fit_transform(X)
labels = Birch(n_clusters=3, threshold=0.5).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.
