[R-sig-Geo] Question about space-time analysis routines

[Nagle, Nicholas]

Corey,

You might also look at Simon Wood's mgcv package.  It is a general purpose Generalized Additive Model package.  There is a way to specify a Conditional Autoregressive Model in space.  And you can interact this with temporal autocorrelation (or equivalently, with temporal splines).  And, since it is based on generalized linear modeling, You can easily specify Poisson type models, which might be important to you in disease rate modeling.

Of course, the problem with this much flexibility is that you can easily create poorly specified models.

Nicholas Nagle

Assistant Professor
Department of Geography
University of Tennessee, Knoxville

On Nov 2, 2012, at 7:07 AM, "r-sig-geo-request at r-project.org" <r-sig-geo-request at r-project.org> wrote:

> Question about space-time analysis routines