Manual Calculation of Average Standard Error for IDW

[JenniferCarroll]

Hello,

I have a data set of about 500 points with depth values. Literature suggests that IDW and kriging are best for bathymetry and so I'm doing a comparison between different models to determine the "best fit" one. As best I understand it, the Geostatistical Analyst does not compute the Average Standard Error, Root-mean-square Standardized, and Mean Standardized errors for IDW because IDW does not include a semivariogram in the modeling process. My question is, is there any way to calculate these errors by hand for an IDW prediction surface?

Thanks!
Jenni

[EricKrause]

IDW cannot predict standard errors because it is not a statistical model.  You can only predict standard errors (and, hence, average standard error) for statistical models like kriging.

[SteveLynch]

Jenni

You may want to add TopoToRaster to your list of interpolation techniques. I'm pretty sure that it'll out perform IDW.

You may also want to look at the SubsetFeatures tool in the Geostatistical Analyst toolbox. With this you can "hide" some data, interpolate and then compare your predictions to the "hidden" values. I.e. model validation.

Regards
Steve

[JenniferCarroll]

Thank you both for your answers. Now that you put it that way, it does make sense for kriging to have the errors and IDW to not have them.

I will also look into the TopotoRaster tool. My professor recommended I use cross validation on my data, but I'll play around with the subset option to see if I get better results and validations. Thanks so much for your help.