Meshed Gaussian Processes

Peruzzi M, Banerjee S, Finley AO (2022) Highly Scalable Bayesian Geostatistical Modeling via Meshed Gaussian Processes on Partitioned Domains. Journal of the American Statistical Association 117(538):969–982.

We introduce a class of models based on the idea that if geospatial data are messy then a Directed Acyclic Graph (DAG) can be used on a tessellation of the spatial domain to create patterns that make things go faster.

Try it yourself using the Shiny app in the meshed R package page.

Data application: Spatially-varying coefficients model on Normalized Difference Vegetation Index (NDVI) on 16 million locations in space & time over the Serengeti park region.