[R-sig-Geo] measuring map similarity?

[Barry Rowlingson]

On Wed, Aug 1, 2012 at 2:39 PM, Jason Robinson
<jason.lesley.robinson at gmail.com> wrote:
> Hoping someone might be able to steer me in the right direction, not sure
> where to look
>
> I have several different modeling methods to predict patterns of species
> occurrences over the same geographic area.  I'd like to compare the shape
> and position of the predicted areas between maps, in reference to a "true
> distribution" map.  Since I have irregular polygons instead of raster cells
> this gets a little hairy for me to think about coding.
>
> I thought that epidemiologists might have some experience with this
> problem.

 Sounds like you could just use the polygon overlay methods in the
rgeos package. Should be trivial to overlay your estimates, A, on your
truth, B, and get three parts - A on its own, B on its own, and the
overlap of A and B.

 gIntersection and gDifference are probably all you need...

Barry