LightGBM
Microsoft's gradient boosting framework optimized for speed and memory efficiency
Microsoft's gradient boosting framework optimized for speed and memory efficiency.
When to use:
- Large datasets (>10k rows)
- Many features
- Need fast training
- Limited memory
Strengths: Very fast, low memory, handles large datasets, accurate Weaknesses: Can overfit small datasets, many hyperparameters
Model Parameters
Num Leaves (default: 31) Maximum number of leaves in one tree. More = complex.
Learning Rate (default: 0.1) Step size for weight updates.
N Estimators (default: 100) Number of boosting iterations.
Max Depth (default: -1) Maximum tree depth (-1 = unlimited).
Feature Fraction (default: 1.0) Fraction of features to use per iteration.
Bagging Fraction (default: 1.0) Fraction of data to use per iteration.
Min Data in Leaf (default: 20) Minimum samples in one leaf.
Reg Alpha, Reg Lambda L1 and L2 regularization.
Random State (default: 42) Seed for reproducibility.