Skip to content
Prev 3664 / 29559 Next

Ordinary and Regression Kriging combined to deal with missing values in predictor variables

Thank you Paul,
I did indeed get more advice from Tomislav Hengl, the same as yours. And 
given the reliability of the (through kriging) predictions of the auxiliary 
variable it indeed makes most sense as you advice also.
Cheers,
Eelke

Tomislav Hengl writes:
I would rather try to extrapolate the predictors (pred1 etc.) and then run
only one model. At least the transition of the values in the final map
will be continuous.

In any case, RK predictions where you do not have values of predictors
will be poor. Only if the regression model is not really significant
(R-square<0.3), then you might rely on the kriging part.

----- Original Message ----- 
From: "Paul Hiemstra" <p.hiemstra at geo.uu.nl>
To: "Eelke Folmer" <E.O.Folmer at rug.nl>
Cc: <r-sig-geo at stat.math.ethz.ch>
Sent: Monday, May 26, 2008 10:47 AM
Subject: Re: [R-sig-Geo] Ordinary and Regression Kriging combined to deal 
with missing values in predictor variables