This paper introduces a probabilistic method for modeling (geo-)spatial tesselations. We propose a new architecture of conditional random fields in two dimensions to estiamte the determinants of spatial groupings.
R programmer; ML enthusiast; Bayesian convert.