Domain Decomposition Approach for Fast Gaussian Process Regression of Large Spatial Data Sets

Gaussian process regression is a flexible and powerful tool for machine learning, but the high computational complexity hinders its broader applications. In this paper, we propose a new approach fo...