LightGBM

LightGBM (Light Gradient Boosting Machine) is a fast, distributed, high-performance gradient boosting framework based on decision tree algorithms (Ke et al., 2017). It is optimized for efficiency and scalability, particularly for large feature spaces, while retaining the accuracy and flexibility of traditional gradient boosting.

Key Features

Extremely fast training and low memory usage.
Handles large datasets and high-dimensional feature spaces efficiently.
Supports categorical features natively.
Implements leaf-wise tree growth for better accuracy.
Compatible with popular ML frameworks and interfaces like scikit-learn.

Official Documentation

LightGBM Documentation

Notes

In this study, LightGBM was evaluated for its speed and efficiency, providing a scalable alternative to XGBoost while maintaining strong predictive performance on tabular datasets.