Results of this run
Growing Neural Gas on Customer Segments grows to maxNodes 10. Gas silhouette is 0.176, gas Calinski-Harabasz is 176.2, gas Davies-Bouldin is 1.622. Held-out segment versus gas units: adjusted Rand 0.412. 0 noise rows out of 200 far from a unit. Gas unit 0 on Customer Segments holds 17 rows (annual_income 80119.4 / avg_spend 2854.5). Gas unit 1 on Customer Segments holds 27 rows (annual_income 26162.2 / avg_spend 737.8). Gas unit 2 on Customer Segments holds 19 rows (annual_income 48674.6 / avg_spend 1640.8). Gas unit 3 on Customer Segments holds 30 rows (annual_income 47894.7 / avg_spend 588.0). Gas unit 4 on Customer Segments holds 22 rows (annual_income 62473.0 / avg_spend 1795.4). Gas unit 5 on Customer Segments holds 22 rows (annual_income 34089.4 / avg_spend 731.1). Gas unit 6 on Customer Segments holds 18 rows (annual_income 80399.6 / avg_spend 2805.8). Gas unit 7 on Customer Segments holds 20 rows (annual_income 40255.8 / avg_spend 598.8). Gas unit 8 on Customer Segments holds 10 rows (annual_income 54746.1 / avg_spend 1232.2). Gas unit 9 on Customer Segments holds 15 rows (annual_income 74055.7 / avg_spend 3243.2). Node-budget path: max_nodes 6 grows 0.257 with 6 groups on Customer Segments; max_nodes 10 grows 0.176 with 10 groups on Customer Segments; max_nodes 16 grows 0.152 with 16 groups on Customer Segments. Gas scatter keeps 90.5%.

| max_nodes | Silhouette | Clusters |
|---|---|---|
| 6 | 0.257 | 6 |
| 10 | 0.176 | 10 |
| 16 | 0.152 | 16 |
# Product Growing Neural Gas: max_nodes=10, random_state=42.
# Features are StandardScaler-transformed before the fit.
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: Synthetic customer records bundled with this site. License: Generated for demos; no external license. The segment 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.
