[R-sig-Geo] moran's I correlogram

[Roger Bivand]

On Wed, 17 Oct 2007, Jennifer Seavey wrote:

> I am looking for R packages that create a moran's I correlogram with 
> ordinal data. my data format is x , y coordinates with a z values that 
> ranges between 0-4. another concern i have is that my sample size is 250 
> and over 60% have a z value of 0. So i am concerned that the uneven 
> nature of the z values will show autocorrelation when it is simply a 
> matter that most of the data points have this value. Your advice/ 
> thoughts are appreciated. Thanks Jenn

Have you considered using join counts (joincount.multi() in spdep)? This 
avoids using the misleading metric in moran.test, but does not use the 
fact that a 3/4 join says more than a 0/4 join. This would have to be 
covered in the description of the output figure. Constructing weights 
lists for successive distance bands, it should be possible to plot a join 
count "correlogram".

Roger

>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

-- 
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