Clustering Algorithm Runner Skill for Comparing K-Means and DBSCAN

Use Claude Code to compare grouping and noise behavior on an interpretable unlabeled dataset.

Install in Claude Code

Use Python and scikit-learn with fixed-seed two-dimensional fixtures. Keep scales comparable or record the scaling rule.

Get the complete `plugins/ai-ml/clustering-algorithm-runner/skills/running-clustering-algorithms` directory from the pinned GitHub commit.
Copy it into the target project's `.claude/skills/`. Check `.claude/skills/running-clustering-algorithms/SKILL.md`, start Claude Code at the project root, and call `/running-clustering-algorithms`.

Example task

Generate 50 points around each of (-4,0) and (4,0) plus three distant points. Fit K-means with k=2 and DBSCAN, then record eps, min_samples, labels, and noise counts.

What to check

Produce a scatter plot with clusters and noise labels. K-means assigns every point, while actual DBSCAN noise counts depend on the parameters and fixture.

Original SKILL.md on GitHub · Updated 2026-11-13

Frequently asked question

Do both algorithms need to assign identical labels?

Compare the resulting shapes and noise behavior. K-means assigns every point, while DBSCAN can mark sparse points as noise.