A Bayesian approach for spatial cluster detection of regression coefficients.
- 2019-06-24 (Mon.), 11:00 AM
- R6005, Research Center for Environmental Changes Building
- Prof. Huiyan Sang
- Department of Statistics, Texas A&M University, USA.
Abstract
In this work, we propose a new Bayesian spatially clustered coefficient (BSCC) regression model, to detect spatial clustering patterns in the associations between response variables and covariates.? In BSCC, regression coefficients are assumed to be constants within each spatially contiguous cluster. To model the clustering patterns, we develop a novel and flexible space partitioning prior to capture irregularly shaped clusters. An efficient Reversible Jump Markov chain Monte Carlo (MCMC) algorithm is designed to estimate the clustered coefficient values and their uncertainty measures. In addition, we provide the Bayesian posterior concentration properties of the estimator. Finally, the performance of the model is illustrated with simulation studies and a real data analysis of temperature-salinity relationship in the Atlantic Ocean.