Abstract

Complex data and modelling may involve different observation mechanisms and resolutions, all tied to a joint collection of latent variables and processes. The inlabru/INLA R package framework allows specification of Bayesian GAMs for such situations, including mild non-linear connections between the latent variables, and combinations of point referenced, aggregated data, and point pattern observations. The general framework has been used on a wide range of applications, including statistical ecology, animal movement, climate research, and earthquake forecasting.