XGBoost: A Scalable Tree Boosting System

Tree boosting is a highly effective and widely used machine learning method. In this paper, we describe a scalable end-to-end tree boosting system called XGBoost, which is used widely by data scientists to achieve state-of-the-art results on many machine learning challenges. We propose a novel sparsity-aware algorithm for sparse data and weighted quantile sketch for approximate tree learning. More

[satojkovic]

detail

[elu_18]

XGBoost: A Scalable Tree Boosting System （´-`）.｡oO( arxivにxgboost論文が… ) https://t.co/Ay0D8x2QUd