XGBoost Article Swipe

PDF

Related Concepts

No concepts available.

Tianqi Chen , Carlos Guestrin ·

YOU? · · 2016 · Open Access · · DOI: https://doi.org/10.1145/2939672.2939785 · OA: W2295598076

Tree boosting is a highly effective and widely used machine learning method.\nIn this paper, we describe a scalable end-to-end tree boosting system called\nXGBoost, which is used widely by data scientists to achieve state-of-the-art\nresults on many machine learning challenges. We propose a novel sparsity-aware\nalgorithm for sparse data and weighted quantile sketch for approximate tree\nlearning. More importantly, we provide insights on cache access patterns, data\ncompression and sharding to build a scalable tree boosting system. By combining\nthese insights, XGBoost scales beyond billions of examples using far fewer\nresources than existing systems.\n