scikit-learn Machine Learning Assistant

What it does

Gives Claude authoritative guidance for classical machine learning with scikit-learn.

• Supervised learning: linear models, SVMs, trees & ensembles, KNN, Naive Bayes, plus an algorithm selection guide
• Unsupervised learning: K-Means, DBSCAN, Agglomerative, Gaussian Mixtures, PCA and manifold methods
• Model evaluation: cross-validation strategies, GridSearchCV/RandomizedSearchCV, classification/regression/clustering metrics, learning and validation curves
• Preprocessing: scaling, categorical encoding, imputation, feature selection, custom transformers
• Pipelines: Pipeline, ColumnTransformer, FeatureUnion for leakage-free workflows

It also ships runnable example scripts (classification pipeline, clustering analysis) and deep-dive reference docs under references/.

Who it's for

• Data analysts and scientists modeling tabular data
• Practitioners who have been bitten by data leakage from preprocessing outside a pipeline
• Researchers needing reproducible ML experiment code for papers or reports
• Beginner-to-intermediate ML users comparing algorithms to establish a baseline

Example uses

1. "Build a churn classification pipeline for my mixed numeric/categorical data and tune it with GridSearchCV" → imputers, one-hot encoding and the estimator wired into one Pipeline
2. "My labels are imbalanced — which metrics should I report?" → guidance on Precision/Recall, ROC AUC, Balanced Accuracy plus a custom scorer example
3. "I don't know how many clusters to use" → elbow and silhouette analysis, K-Means vs DBSCAN vs GMM comparison, and a PCA projection plot

Install the skill I found on Claude Skill Mart.
Copy the skills/scikit-learn folder from the GitHub repo K-Dense-AI/scientific-agent-skills into my ~/.claude/skills/scikit-learn/.
When it's done, tell me in one line what this skill can do.