[GRASS-dev] PIxel-wise regression contribution

[Damien O'Grady]

Hi there

I would like to contribute a script that I have used in my research to
the broader field.  The script carries out linear regression, except
that rather than using two rasters for the x and y variables, it uses
any number of pairs of rasters, allowing the regression to be carried
out across a time series, resulting in a slope, intercept and R^2 values
for each pixel.

The crunch is that it is written in perl, as this is what I know best
and it serves me well.  I have done my best to conform to the stated
standards, and I have produced a fairly detailed manual page.  This will
give you a better idea of what the script does, with examples.  I attach
the script and html here.  I hope that you will consider making it
available to others.

Here at James Cook University, by the way, we are encouraging a switch
from problematic and costly proprietary software to GRASS GIS.  I am a
major fan of GRASS, particularly on a Linux platform, and I am pleased
that more and more students are listening.  You should see a sharp rise
in its use here in Queensland in the near future. 

Thanks

Damien O'Grady

-------------- next part --------------
A non-text attachment was scrubbed...
Name: ogrady.tar.gz
Type: application/x-compressed-tar
Size: 5234 bytes
Desc: not available
Url : http://lists.osgeo.org/pipermail/grass-dev/attachments/20120519/842897c3/ogrady.tar.bin