Hello List, I am trying to find an R package that will accommodate spatially constrained clustering. While I have been unable to find a package that is explicitly designed to do spatially constrained clustering, I was wondering if anyone had found a package that would do constrained clustering of any kind, and adapted this to spatial constraints? I have searched the R site extensively, and googled all night long, but to no avail! I HAVE found this post: http://finzi.psych.upenn.edu/R/Rhelp02a/archive/56819.html but the replies did not help much. They lead to several packages which perform spatial clustering (such that significant clusters of say a disease are located within a study region), however, what I would like to do is partition a spatial (grid) dataset based on multiple variables, taking into account their spatial locations (i.e. clustering is based on the variables, but constrained so that clusters are spatially contiguous). I'm thinking mclust is probably the best way to go, but I'm not sure where to start. Any suggestions would be greatly appreciated. Thanks, Carson
Spatially Constrained Clustering
3 messages · Carson Farmer, Gavin Simpson, Elias T. Krainski
Carson, If you don't get any other firm answers, take a look at this work in progress of Steve Juggins: http://www.campus.ncl.ac.uk/staff/Stephen.Juggins/analysis.htm He hacked the source of hclust (and the underlying Fortran) to do constrained clustering using hclust's methods. This was for a single constraint so we can keep samples in time together, but I'm sure by studying the code, if you know Fortran (and I don't) and compare the minor changes Steve made to the code for hclust, you might be able to hack it to do what you want. HTH G
On Mon, 2008-01-14 at 15:42 -0800, Carson Farmer wrote:
Hello List, I am trying to find an R package that will accommodate spatially constrained clustering. While I have been unable to find a package that is explicitly designed to do spatially constrained clustering, I was wondering if anyone had found a package that would do constrained clustering of any kind, and adapted this to spatial constraints? I have searched the R site extensively, and googled all night long, but to no avail! I HAVE found this post: http://finzi.psych.upenn.edu/R/Rhelp02a/archive/56819.html but the replies did not help much. They lead to several packages which perform spatial clustering (such that significant clusters of say a disease are located within a study region), however, what I would like to do is partition a spatial (grid) dataset based on multiple variables, taking into account their spatial locations (i.e. clustering is based on the variables, but constrained so that clusters are spatially contiguous). I'm thinking mclust is probably the best way to go, but I'm not sure where to start. Any suggestions would be greatly appreciated. Thanks, Carson
_______________________________________________ R-sig-Geo mailing list R-sig-Geo at stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/r-sig-geo
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~% Dr. Gavin Simpson [t] +44 (0)20 7679 0522 ECRC, UCL Geography, [f] +44 (0)20 7679 0565 Pearson Building, [e] gavin.simpsonATNOSPAMucl.ac.uk Gower Street, London [w] http://www.ucl.ac.uk/~ucfagls/ UK. WC1E 6BT. [w] http://www.freshwaters.org.uk %~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
Hello Carson, See the SKATER software at http://www.est.ufmg.br/leste/skater.htm The SKATER is a Spatial 'K'luster Analisys by Tree Edge Removal. In future, this method also be available in R. Best, Elias. --- Carson Farmer <cfarmer at uvic.ca> escreveu:
Hello List, I am trying to find an R package that will accommodate spatially constrained clustering. While I have been unable to find a package that is explicitly designed to do spatially constrained clustering, I was wondering if anyone had found a package that would do constrained clustering of any kind, and adapted this to spatial constraints? I have searched the R site extensively, and googled all night long, but to no avail! I HAVE found this post:
http://finzi.psych.upenn.edu/R/Rhelp02a/archive/56819.html
but the replies did not help much. They lead to several packages which perform spatial clustering (such that significant clusters of say a disease are located within a study region), however, what I would like to do is partition a spatial (grid) dataset based on multiple variables, taking into account their spatial locations (i.e. clustering is based on the variables, but constrained so that clusters are spatially contiguous). I'm thinking mclust is probably the best way to go, but I'm not sure where to start. Any suggestions would be greatly appreciated. Thanks, Carson
_______________________________________________ R-sig-Geo mailing list R-sig-Geo at stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Elias T. Krainski