Articles | Volume 10, issue 2
https://doi.org/10.5194/ascmo-10-143-2024
https://doi.org/10.5194/ascmo-10-143-2024
29 Oct 2024
 | 29 Oct 2024

A robust approach to Gaussian process implementation

Juliette Mukangango, Amanda Muyskens, and Benjamin W. Priest

Model code and software

Juliettengango1/Robust-GP-Implementation: Robust MuyGPyS (Version v1) Juliette Mukangango https://doi.org/10.5281/zenodo.13987737

Download
Short summary
In this study, we investigated the performance of Gaussian process regression (GP) models in handling outlier-affected spatial datasets. Our findings emphasized that models with the proposed methods provided accurate predictions and reliable uncertainty quantification, showcasing resilience against outliers. Overall, our study contributes to advancing the understanding of GP regression in spatial contexts and offers practical solutions to enhance its applicability in outlier-rich environments.