Learning non Gaussian spatial distributions via Bayesian transport maps with parametric shrinkage
This paper proposes a novel approach, ShrinkTM, for modelling high-dimensional, non-Gaussian spatial distributions using Bayesian transport maps. ShrinkTM improves upon existing methods by introducing a shrinkage mechanism that pushes the map components towards a parametric Gaussian family, resulting in more accurate inference, particularly when only a limited number of training samples are available. The paper demonstrates the effectiveness of ShrinkTM through numerical experiments on simulated data and climate-model output, showcasing its ability to capture complex dependencies and outperform existing methods, even with a single training sample. https://arxiv.org/abs/2409.19208
Significance
Section titled “Significance”Understanding these findings helps advance our knowledge and inform better decisions. This research represents an important contribution to the field. For the full details, watch the video above and explore the linked resources.
Youtube Hashtags
Section titled “Youtube Hashtags”#machinelearning #bayesian #spatialdata #datascience #statistics
Youtube Keywords
Section titled “Youtube Keywords”learning non gaussian spatial distributions via bayesian transport maps with parametric shrinkage
ResearchLounge
https://researchlounge.org/interdisciplinary-fields/climate-impact/learning-non-gaussian-spatial-distributions-via-bayesian-transport-maps-with-parametric-shrinkage/