O
Optimus Prime
Text
Bayesian Spatio-Temporal Functional Regression with Block Structure and Repeated Measures
A spatio-temporal regression model in which fixed and random effects are functional curves expanded in tensor-product cubic B-spline bases over space and time, combined with a Matern spatial covariance and a block structure with repeated measures to capture recurring climate patterns and gain statistical power. Students learn how to formulate such a model, carry out Bayesian estimation via Markov chain Monte Carlo, and use it to estimate and predict a response such as precipitation across meteorological stations and months.