[R-sig-Geo] question about regression kriging

Tue Apr 8 15:56:54 CEST 2008

Hi,

Tom and Thierry, Thank you for your advice, the lecture notes are very useful. We will try geoRglm but for now regression kriging using the working residuals gives sensible answers even though there are some issues with using working residuals, i.e. not Normally distributed, occasional very large values and inv.logit(prediction type="link" + working residual) doesn't quite give the observed values.

Our final question about this is how to estimate standard errors for the regression kriging predictions of the binary variable?

On the logit scale we are using
 rk prediction (s0) = glm prediction (s0) + kriged residual prediction (s0) 
for location s0

Is assuming independence of the two components adequate?
 var rk(s0) ~= var glm prediction (s0) + var kriged residual prediction (s0) 

Thanks again,
David Maxwell & Vanessa Stelzenmüller

david.maxwell at cefas.co.uk

-----Original Message-----
From: r-sig-geo-bounces at stat.math.ethz.ch [mailto:r-sig-geo-bounces at stat.math.ethz.ch] On Behalf Of Tomislav Hengl
Sent: 03 April 2008 09:04
To: r-sig-geo at stat.math.ethz.ch
Subject: Re: [R-sig-Geo] question about regression kriging

I completely agree with Thierry.

Take a look at this also:
https://stat.ethz.ch/pipermail/r-sig-geo/2008-February/003176.html 

The instructions on how to run RK with binary variables in R you can find in sec 4.3.3 (Fig. 4.15)
of my lecture notes.

Hengl, T., 2007. A Practical Guide to Geostatistical Mapping of Environmental Variables. EUR 22904
EN Scientific and Technical Research series, Office for Official Publications of the European
Communities, Luxemburg, 143 pp.
http://bookshop.europa.eu/uri?target=EUB:NOTICE:LBNA22904:EN:HTML 

Tom Hengl
http://spatial-analyst.net 

-----Original Message-----
From: r-sig-geo-bounces at stat.math.ethz.ch [mailto:r-sig-geo-bounces at stat.math.ethz.ch] On Behalf Of
ONKELINX, Thierry
Sent: woensdag 2 april 2008 11:56
To: Vanessa Stelzenmuller (Cefas); r-sig-geo at stat.math.ethz.ch
Subject: Re: [R-sig-Geo] question about regression kriging

Dear Vanessa,

What residuals did you use? The ones in the original scale or in the logit scale? Interpolate the
residuals in the logit scale and add these to the model predictions in the logit scale. And the
transform those values back to the original scale. This will prevent values outside the 0-1 range.

Maybe you should have a loot at the geoRglm package.

HTH,

Thierry

----------------------------------------------------------------------------
ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature and Forest
Cel biometrie, methodologie en kwaliteitszorg / Section biometrics, methodology and quality
assurance
Gaverstraat 4
9500 Geraardsbergen
Belgium 
tel. + 32 54/436 185
Thierry.Onkelinx at inbo.be 
www.inbo.be 

To call in the statistician after the experiment is done may be no more than asking him to perform a
post-mortem examination: he may be able to say what the experiment died of.
~ Sir Ronald Aylmer Fisher

The plural of anecdote is not data.
~ Roger Brinner

The combination of some data and an aching desire for an answer does not ensure that a reasonable
answer can be extracted from a given body of data.
~ John Tukey

-----Oorspronkelijk bericht-----
Van: r-sig-geo-bounces at stat.math.ethz.ch [mailto:r-sig-geo-bounces at stat.math.ethz.ch] Namens Vanessa
Stelzenmuller (Cefas)
Verzonden: woensdag 2 april 2008 11:32
Aan: r-sig-geo at stat.math.ethz.ch
Onderwerp: [R-sig-Geo] question about regression kriging

Hello,

We work on the application of regression kriging to presence / absence data in the context of
species distribution modelling. In R in a first step we fit the trend surfaces with logistic
regression models. Then we fit a variogram to the regression residuals and interpolate the residuals
with OK. Now we face the situation that when combining trend surfaces with residual surfaces for
some locations our occurrence probability is <0 or >1. Thus taking into account the spatial
structure of the data (residuals)  has the potential to convert a predicted high occurrence
probability into a low occurrence probability or vice versa. Are there some restriction for
presence/ absence data for this approach? How to deal with these estimations (<0 and >1)?

Many thanks

Vanessa  

________________________________

Dr. Vanessa Stelzenmüller

Marine Scientist (GIS), CEFAS

Pakefield Road, Lowestoft, NR33 0HT, UK

Tel.: +44 (0)1502 527779

www.cefas.co.uk

***********************************************************************************
This email and any attachments are intended for the =\...{{dropped:20}}