The most common cause is that you have multiple observations at single
locations. You can use function zerodist to figure out whether this is
the case.
On 10/14/2011 06:11 PM, Sandeep Patil wrote:
Hello
I am running a Kriging cross validation in Automap. I get the following
error
Hello all
I have about 4500 spatial points which are irregularly spaced...
What is want to do is to replace points that are clustered very near
each other by one point ...i.e say the distance betweeen Point A Point B
is less than a threshold that i will define , i will delete either A or B
Hello List,
I have a map of Africa I plot using spplot(). In addition, I have a second map
of Sao Tome (a small island). I want to add this second map to the first as an
inset because it would otherwise not be visible at the same scale of the
continent.
This is what I currently do: I save
As you gave practically no information besides the error message, why
don't you simply try that.
Another possible cause is using a Gaussian variogram model with zero nugget.
On 10/15/2011 04:48 PM, Sandeep Patil wrote:
I did check the data once again and did not find any such points .
This doesn't seem to be a permission issue, as there are several other
compiled modules in the R library directory. It seems pretty clear that
either libgdal.so.1 is pointing to the wrong place, or does not have
references to PROJ.4.
Any ideas? Any hints? I could sure use the extra computing
you might want to look at maptools::elide, to shift Sao to the Africa
map, and perhaps rbind() the two objects into a single object for spplot.
Alternative, pass Sao in sp.layout, but than you'd have to do the
colouring yourself, which might be fine if you need to do it just once.
On
Sandeep,
I had a similar problem when I worked with wildlife GPS data and cluster
analysis, and I had over 50,000 points.
What I essentially did was... create a distance table for all points
(4500x4500 in your case)
Every row represents a spatial point and every column represents every other