LightGBM
LightGBM is one of the most popular machine learning algorithms for building predictive models for both classification and regression tasks. It belongs to the same gradient boosting family as XGBoost, while achieving faster training and better memory efficiency.
For example, techniques such as the "Leaf-wise" strategy that prioritizes splitting only the leaf that most improves prediction accuracy, GOSS that skips data already being predicted well, and EFB that bundles features that don't take values at the same time, enable efficient model building even on large-scale data.
For more details on how to use it, please refer to the following articles.
Last updated