ideq: Bayesian Dynamic Spatio-Temporal Models, Including the Integrodifference Equation Model

In contrast to other methods of modeling spatio-temporal data, dynamic spatio-temporal models (DSTMs) directly model the dynamic data-generating process. 'ideq' supports two main classes of DSTMs: (1) empirical orthogonal function (EOF) models and (2) integrodifference equation (IDE) models. EOF models do not directly use any spatial information; instead, they make use of observed relationships in the data (the principal components) to model the underlying process. In contrast, IDE models are based on diffusion dynamics and the process evolution is governed by a (typically Gaussian) redistribution kernel. Both types have a variety of options for specifying the model components, including the process matrix, process error, and observation error. The classic reference for DSTMs is Noel Cressie and Christopher K. Wikle (2011, ISBN:978-0471692744). For IDE models specifically, see Christopher K. Wikle and Noel Cressie (1999, <https://www.jstor.org/stable/2673587>) and Christopher K. Wikle (2002, <doi:10.1191/1471082x02st036oa>).

Downloads:

Linking:

Please use the canonical form https://CRAN.R-project.org/package=ideq to link to this page.