[R] Tps error

From: Jennifer Gremer <jgremer_at_email.arizona.edu>
Date: Fri, 13 May 2011 11:13:08 -0700

I am trying to use the Tps procedure to visualize 3 variables in XYZ space, but keep getting the error messages below:
Warning messages:
1: In Krig.find.gcvmin(info, lambda.grid, gcv.grid$GCV, Krig.fgcv, :   GCV search gives a minimum at the endpoints of the grid search
2: In Krig.find.gcvmin(info, lambda.grid, gcv.grid$GCV.one, Krig.fgcv.one, :   GCV search gives a minimum at the endpoints of the grid search

This is the code I am using:
tpsfit <- Tps(cbind(x1, x2), Y, scale.type = "unscaled")

I have successfully used the Tps function for other subsets of the data, but one set is problematic. The output suggests that the smoothing parameter (lambda) is very low (5.632e-05). I have also attached the rest of the output from the function below this email.

Can anyone help me with this problem? I can send an example of the data I am using if that is helpful.

Any feedback would be appreciated.
Thank you,

Tps(x = cbind(x1, x2), Y = Y, scale.type = "unscaled", cov.function = "Thin plate spline radial basis functions (Rad.cov) ")

 Number of Observations:                116
 Number of unique points:               116
 Number of parameters in the null space 3
 Parameters for fixed spatial drift     3
 Effective degrees of freedom:          110.2
 Residual degrees of freedom:           5.8
 MLE sigma                              0.03841
 GCV sigma                              0.05004
 MLE rho                                26.2
 Scale passed for covariance (rho)      <NA>
 Scale passed for nugget (sigma^2)      <NA>
 Smoothing parameter lambda             5.632e-05

Residual Summary:
       min      1st Q     median      3rd Q        max
-0.0339500 -0.0054520 0.0002131 0.0044340 0.0351400

Covariance Model: Rad.cov
  Names of non-default covariance arguments:


 Method used: GCV Cost: 1

   lambda       trA       GCV   GCV.one GCV.model      shat
5.632e-05 1.102e+02 5.007e-02 5.007e-02        NA 5.004e-02

 Summary of all estimates found for lambda
              lambda     trA     GCV    shat   REML
GCV        5.632e-05 110.200 0.05007 0.05004 40.029
GCV.model         NA      NA      NA      NA     NA
GCV.one    5.632e-05 110.200 0.05007 0.05004     NA
RMSE              NA      NA      NA      NA     NA
pure error        NA      NA      NA      NA     NA
REML       8.520e+03   3.001 0.06374 0.24919  3.323

Jennifer R. Gremer
Postdoctoral Research Associate
Ecology and Evolutionary Biology
University of Arizona
mail: P.O. 210088
deliveries: Biosciences West, Room 310
Tucson, AZ 85721
office: 520-621-4022

R-help_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Received on Fri 13 May 2011 - 20:53:24 GMT

This quarter's messages: by month, or sorted: [ by date ] [ by thread ] [ by subject ] [ by author ]

All messages

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Fri 13 May 2011 - 21:20:06 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive