# Xgboost

> XGBoost gradient boosting library. Use for tabular ML.

- Skill: `majiayu000/xgboost` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add majiayu000/xgboost`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/xgboost/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/xgboost

---


# XGBoost

XGBoost is the winningest algorithm in Kaggle history for tabular data. v2.1 (2025) brings native **Blackwell** GPU support and Polars integration.

## When to Use

- **Tabular Data**: It usually beats Deep Learning on structured tables.
- **Speed**: Extremely optimized C++ backend.

## Core Concepts

### Gradient Boosting

Building extensive decision trees sequentially, each correcting the previous one's errors.

### DMatrix

Internal optimized data structure.

### Device Parameter

`device="cuda"` enables GPU acceleration.

## Best Practices (2025)

**Do**:

- **Use `device="cuda"`**: GPU training is 10x faster.
- **Use Early Stopping**: Stop training when validation error rises.
- **Pass Polars Dataframes**: No need to convert to Pandas/NumPy first.

**Don't**:

- **Don't use one-hot encoding**: Use native categorical support (`enable_categorical=True`).

## References

- [XGBoost Documentation](https://xgboost.readthedocs.io/)