Skip to main content

> ML_ALGORITHM // GRADIENT-BOOSTING_v1.0

Gradient Boosted Decision Trees (GBDT)

Sequential ensemble technique that trains shallow trees to predict pseudo-residuals of previous iterations along the gradient of the loss.

Tree & Rule Ensemblesclassical-supervisedmoderate-posthocmedium (1k-100k)
Back to All Algorithms
Computational Complexity
Training Complexity:O(n_trees * p * n log n)
Inference Complexity:O(n_trees * depth)
Hardware Profile
CPU Friendly:Yes
Requires GPU:No
Memory Footprint:moderate
Interpretability & Data
Interpretability Tier:moderate-posthoc
Training Data Needs:medium (1k-100k)

Interpretability Assessment

TreeSHAP provides mathematically rigorous local and global feature attributions.

Suitable Tasks & Supported Modalities

Suitable Tasks:
binary classificationmulticlass classificationregressionranking
Supported Modalities:
tabular

Implementing Libraries

scikit-learnscikit-learn Consortium / Inria · v1.5.2
View Spec
caret-r

Foundational Literature

Common Pitfalls & Warnings
  • Overfitting without early stopping and shrinkage (learning rate)
  • Inability to extrapolate linear trends outside training feature boundaries