Dear list,
Having fit a variogram to a dataset (indoor radon measurements) and
applied cross-validations, I noticed the perfect negative correlation
(-0.95) between my kriging residuals and my input data.
This means that I am overestimating as much the low values as I am
underestimating the
Gregoire and List-
Sorry I didn't read bottom of email (phone screen is too small)
Pure nugget means estimate is mean of data
Correlation between (Mean-Data) vs Data is -1
I assume -.95 is because your search radius doesn't use all the data or
excluding the 1 point in xvalidation
Hi Edzer,
I thought about that (the nugget effect) but my variogram (and robust
variograms) did show a short scale structure (with a high nugget
effect).
Considering the possibility you mention, that I actually have a
phenomenon with a pure nugget effect although I see a structure, what do
I
Gregoire-
Sounds like white noise to me and your estimate is the mean eg 3 point
example
Input -1 0 1
Est 0 0 0
Res 1 0 -1 (as est-input)
Is your variogram high nugget?
Robert (Bob) L. Sandefur PE
Senior Geostatistician / Reserve Analyst
CAM
200 Union Suite 430
Hi, Gregoire,
what is the error of your measurements? Perhaps this is masking your data?
sometimes they can be very big!
Are you able to know it?
Hope this helps,
Carme
En/na Gregoire Dubois ha escrit:
Hi Edzer,
I thought about that (the nugget effect) but my variogram (and robust
Gregoire
The correlation between actual value and error of estimation is always
present to some extent and is simply due to the estimation process. High values
will b eunderestimated from neighbouring samples. Low values will be
overestimated from neighbouring samples. The only way you
Dear Gregoire
As one who has dealt extensively with radon (but not from the
statistical world), I would appreciate it if you could post to one of
your sites either histogram in either log or some other plot. (or
your data if possible)
I have been dealing with over 50,000 data points
Hello Gregoire, and list,
I'd guess that, although you may be seeing some spatial structure in your
data, the effective range of your model may be less than the minimum distance
between the locations you are predicting and their closest known locations.
In other words, your model may
Gregoire,
The question of whether or not to count degrees of freedom still does not get
the attention it so richly deserves. Once upon a time I asked a prominent
professor of mathematics whether or not degrees of freedom may be dismissed
with impunity. His response was, But without degrees of
