[R] Monmonier algorithm

Thibaut Jombart jombart at biomserv.univ-lyon1.fr
Mon Aug 27 13:39:15 CEST 2007


Hello,

Here is a late answer, but an answer nonetheless to the question I asked 
almost one year ago on this list:

 > On Wed, 29 Mar 2006, Thibaut Jombart wrote:

 >> Hello list, 
<http://tolstoy.newcastle.edu.au/R/help/06/03/24318.html#24322qlink1>
/>> /
/>> does anyone know if Monmonier algorithm is available in R? I've 
checked /
/>> several spatial libraries, but I didn't find anything related to it. /
/>> However, there is a huge documentation and I may have missed it. /
/>> /
/>> Before coding it, I'd like to be sure it doesn't already exist. /

 > Googling, I found:

 > http://www-med-physik.vu-wien.ac.at/staff/rub/abstracts/ISCB_2005.pdf

 > which is a poster, and refers to using R for boundary finding, and 
other software for data management and display.  >Perhaps the authors 
are able to help by making code available, the poster looks like a nice 
example of spatial data >analysis.

> -- 
> Roger Bivand
> Economic Geography Section, Department of Economics, Norwegian School of
> Economics and Business Administration, Helleveien 30, N-5045 Bergen,
> Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
> e-mail: Roger.Bivand at nhh.no

Basically, Monmonier algorithm aims at finding maximum-difference 
boundaries between geo-referenced objects. It requires a set of 
georeferenced objects along with matrix of distances among these objects.

Monmonier algorithm is now implemented in the adegenet package  
(http://pbil.univ-lyon1.fr/software/adegenet/). Main functions are 
'monmonier' and 'optimize.monmonier'. Despite the package is devoted to 
genetic data analysis, these functions can handle other kind of data as 
well.

The main difference I can see between this implementation and the 
original algorithm is that here, the function uses objects connected on 
a neighbouring graph rather than polygons of a Voronoi tesselation. 
Thus, Delaunay triangulation shall be used to recover the original 
version of the algorithm, but other graphs are also possible (e.g. 
Gabriel's graph).

Regards,

Thibaut.

-- 
######################################
Thibaut JOMBART
CNRS UMR 5558 - Laboratoire de Biométrie et Biologie Evolutive
Universite Lyon 1
43 bd du 11 novembre 1918
69622 Villeurbanne Cedex
Tél. : 04.72.43.29.35
Fax : 04.72.43.13.88
jombart at biomserv.univ-lyon1.fr
http://lbbe.univ-lyon1.fr/-Jombart-Thibault-.html?lang=en



More information about the R-help mailing list