How does this skill help with data leakage?

It identifies preprocessing steps performed outside of cross-validation loops and refactors them into integrated Scikit-learn Pipelines, ensuring transformers are only fitted on training data.

Does it support hyperparameter tuning optimization?

Yes, it implements systematic patterns for GridSearchCV and RandomizedSearchCV, including the correct syntax for addressing parameters within nested pipeline steps.

Can it handle custom feature engineering logic?

Yes, the skill can transform manual feature engineering code into standard Scikit-learn custom transformers using the BaseEstimator and TransformerMixin classes.

How does it ensure ML model reproducibility?

The skill audits your code for missing random_state parameters in estimators and splitters, ensuring that seeds are applied consistently across the entire ML workflow.

Scikit-Learn Refactoring & Optimization

Name: Scikit-Learn Refactoring & Optimization
Author: SnakeO

bySnakeO

0•

数据科学与机器学习

Refactors Scikit-learn and machine learning code into production-ready pipelines that ensure reproducibility and prevent data leakage.

This skill transforms experimental machine learning scripts into robust, maintainable production code by enforcing Scikit-learn best practices. It automatically implements Pipeline and ColumnTransformer architectures, ensures systematic reproducibility through random_state management, and converts manual preprocessing into reusable custom transformers. It is particularly effective at cleaning up fragmented notebook code, preventing data leakage during cross-validation, and ensuring that feature engineering is properly encapsulated for deployment environments.

主要功能

01Standardized hyperparameter tuning and evaluation patterns

02Systematic reproducibility through global random state management

03Automated Pipeline and ColumnTransformer implementation

040 GitHub stars

05Data leakage prevention in preprocessing and cross-validation

06Custom transformer generation following Scikit-learn API standards

使用场景

01Auditing ML code for data leakage and non-reproducible patterns

02Refactoring manual data preprocessing into encapsulated Scikit-learn Pipelines

03Converting Jupyter Notebook experiments into production-grade Python scripts

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add snakeo/claude-debug-and-refactor-skills-plugin refactor-scikit-learn

For use in Claude.ai and ChatGPT

Download Skill