[GRASS-dev] library of data partitioning algorithms

Hello,

In the effort of creating a C-version of d.vect.thematic we are planning on using a series of library functions for data partitioning which would contain classical algorithms such as equal interval, quantiles, stdev, etc, but also more sophisticated ones such as "natural breaks" or kmeans. They would take a vector of data and the number of desired classes and return the breakpoints.

These functions could then be used by other modules, such as d.vect.chart.

I have two questions concerning this:

1) Where should we integrate these functions ? lib/stats is currently limited to raster, should we integrate non-raster functionality in there ?

2) Does anyone know of an existing C-library of such functions. Just so that we don't reinvent the wheel.

Thanks,

Moritz

On Mon, 27 Nov 2006, Moritz Lennert wrote:

Hello,

In the effort of creating a C-version of d.vect.thematic we are planning
on using a series of library functions for data partitioning which would
contain classical algorithms such as equal interval, quantiles, stdev,
etc, but also more sophisticated ones such as "natural breaks" or
kmeans. They would take a vector of data and the number of desired
classes and return the breakpoints.

These functions could then be used by other modules, such as d.vect.chart.

I have two questions concerning this:

1) Where should we integrate these functions ? lib/stats is currently
limited to raster, should we integrate non-raster functionality in there ?

2) Does anyone know of an existing C-library of such functions. Just so
that we don't reinvent the wheel.

You'll find some references in the R classInt package on CRAN, including
the Fortran for Fisher-Jenks (not C). There is a good deal of C code in R
itself, some from statlib.

Roger

Thanks,

Moritz

_______________________________________________
grass-dev mailing list
grass-dev@grass.itc.it
http://grass.itc.it/mailman/listinfo/grass-dev

--
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@nhh.no