XGBoost

XGBoost implements gradient-boosted decision trees with a strong focus on performance and scalability, and has been one of the most consistently winning algorithms in machine learning competitions for structured (tabular) data - the kind of data found in spreadsheets and databases rather than images or text. It runs efficiently on a single machine or distributed across a cluster.

XGBoost is a fit for teams solving classification, regression, or ranking problems on structured data, a different tool than the deep learning frameworks (PyTorch, TensorFlow, also in this catalog) that dominate image, text, and audio tasks but are often outperformed by gradient boosting on tabular data.