skills.rest
KameniAlexNea
KameniAlexNeaCOMMUNITY
Explore all skills by this author →
𝕏@kamenialexnea22GitHub

feature-engineering

Generate and prune numerical, categorical, and temporal features with SHAP-based importance.

Updated Feb 24, 2026
Search Tags:#python#feature engineering#data leakage#shap#feature selection#ml competitions
One-click install
npx skills add https://github.com/KameniAlexNea/gladius-agent --skill feature-engineering-kamenialexnea
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: feature-engineering
Source: https://github.com/KameniAlexNea/gladius-agent/tree/main/gladius/utils/templates/skills/feature-engineering
Command: npx skills add https://github.com/KameniAlexNea/gladius-agent --skill feature-engineering-kamenialexnea

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses the challenge of improving machine learning model performance by systematically generating, evaluating, and selecting relevant features, while rigorously preventing data leakage.

Core Features & Use Cases

  • Leakage Prevention: Provides strict rules and patterns to avoid common data leakage pitfalls.
  • Diverse Feature Recipes: Offers pre-built scripts for numerical, categorical, and temporal feature engineering.
  • SHAP-based Importance & Pruning: Integrates SHAP values for feature importance and a disciplined pruning process to retain only impactful features.
  • Use Case: After establishing a baseline model, use this Skill to generate advanced features like target-encoded categoricals (safely within CV folds), lag features for time-series data, and interaction terms, then prune them based on SHAP importance and OOF score improvement.

Quick Start

Use the feature-engineering skill to add numerical and categorical features to your dataset, then evaluate their importance using SHAP.

Frequently Asked Questions about feature-engineering

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent data leakage when generating features in Python?▼

To prevent data leakage during feature generation, you need strict rules like computing target-encoded categoricals safely within cross-validation folds. This approach ensures that information from validation data does not bleed into training features.

What is the best way to select features using SHAP values?▼

The best way to select features using SHAP values is through an incremental pruning workflow. You measure feature importance with SHAP, retain only impactful features, and verify if OOF scores improve after pruning.

How do I create lag features for time-series data in Python?▼

You create lag features for time-series data by applying pre-built Python recipes designed for temporal feature generation. These scripts systematically build time-based features while maintaining leakage prevention boundaries.

Can I use SHAP-based pruning after building a baseline model?▼

Yes, you can use SHAP-based pruning after establishing a baseline model. You generate advanced features like interaction terms, evaluate their importance, and prune them to optimize your feature set.

What types of features can I generate for ML competitions using Python?▼

For ML competitions, you can generate numerical, categorical, and temporal features using Python scripts. These include target-encoded categoricals, interaction terms, and time-series lag features.

Why should I prune features instead of keeping all generated ones?▼

You should prune features to retain only impactful ones and improve model performance. Using SHAP-based importance measurement, an incremental pruning workflow removes unnecessary features and checks OOF score improvement.

Related Skills

Explore More Data & Analytics Skills→
doanchienthangdevdoanchienthangdev
4

feature-engineering

Transform, select, and manage machine learning features via feature stores.

Community
Advanced
snoodleboot-iosnoodleboot-io
2

cross-validation-strategies

Provides k-fold and bootstrap cross-validation strategies for ML models to detect overfitting and improve generalization on unseen data.

Official
Intermediate
Ritabrata-ChakrabortyRitabrata-Chakraborty

shap

Computes SHAP values to interpret ML model predictions and feature importance.

Community
Advanced
tomlupotomlupo

shap

Explain model predictions by attributing outputs to input features with SHAP values.

Community
Advanced