Patchwork kriging for large-scale Gaussian process regression

Chiwoo Park, Daniel Apley

Research output: Contribution to journalArticlepeer-review

46 Scopus citations

Abstract

This paper presents a new approach for Gaussian process (GP) regression for large datasets. The approach involves partitioning the regression input domain into multiple local regions with a different local GP model fitted in each region. Unlike existing local partitioned GP approaches, we introduce a technique for patching together the local GP models nearly seamlessly to ensure that the local GP models for two neighboring regions produce nearly the same response prediction and prediction error variance on the boundary between the two regions. This largely mitigates the well-known discontinuity problem that degrades the prediction accuracy of existing local partitioned GP methods over regional boundaries. Our main innovation is to represent the continuity conditions as additional pseudo-observations that the differences between neighboring GP responses are identically zero at an appropriately chosen set of boundary input locations. To predict the response at any input location, we simply augment the actual response observations with the pseudo-observations and apply standard GP prediction methods to the augmented data. In contrast to heuristic continuity adjustments, this has an advantage of working within a formal GP framework, so that the GP-based predictive uncertainty quantification remains valid. Our approach also inherits a sparse block-like structure for the sample covariance matrix, which results in computationally efficient closed-form expressions for the predictive mean and variance. In addition, we provide a new spatial partitioning scheme based on a recursive space partitioning along local principal component directions, which makes the proposed approach applicable for regression domains having more than two dimensions. Using three spatial datasets and three higher dimensional datasets, we investigate the numerical performance of the approach and compare it to several state-of-the-art approaches.

Original languageEnglish (US)
Pages (from-to)1-43
Number of pages43
JournalJournal of Machine Learning Research
Volume19
StatePublished - Jul 1 2018

Funding

The authors are thankful for generous support of this work. Park was supported in part by the grants from National Science Foundation (CMMI-1334012) and Air Force Office of Scientific Research (FA9550-18-1-0144). D. Apley was supported in part by National Science Foundation (CMMI-1537641).

Keywords

  • Local Kriging
  • Model Split and Merge
  • Pseudo Observations
  • Spatial Partition

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence
  • Control and Systems Engineering
  • Statistics and Probability

Fingerprint

Dive into the research topics of 'Patchwork kriging for large-scale Gaussian process regression'. Together they form a unique fingerprint.

Cite this