Diffusion Interpolation with Barriers - extent

[BankimYadav]

Hello. 

The extent of the output using the tool 'Diffusion Interpolation with Barriers' is coming wrong. 

I have specified the extent in the environment settings as : 

with arcpy.EnvManager(extent = <some raster>)

or

arcpy.env.extent = <some raster>

The output is not covering the entire raster extent. interpolations geostatistical analys 

[DanPatterson_Retired]

You might want to move this from My Esri‌ to the Geostatistical Analyst‌ place if you need help.

Also, could you provide the parameters that you are using to help anyone that might see it

[EricKrause · Accepted solution]

Hi Bankim,

Are you performing Diffusion Interpolation using the Geostatistical Wizard or the geoprocessing tool?  If it's the Wizard, it will not honor geoprocessing environments.  If you have a layer that you created in the Wizard, and you want to change its extent, you can use the Create Geostatistical Layer geoprocessing tool.  Provide the old layer, provide the datasets used to create the layer, and give the new layer a name.  This tool will honor geoprocessing environments, so you can set the extent environment, and the new layer will be identical to the old layer except with the new extent.

Thanks,

Eric

[BankimYadav]

Oh, so its the master himself.  Great.  I got what you said by using the GA tool for GPI in the software. I was getting wrong results somehow while using the tool through python and arcpy. 

Still, I'd like to ask I used a degree = 9 in the GPI as it was giving the best cross-validation results though I have read somewhere in GA documentation that a degree upto 3 is recommended. Please comment. 

[EricKrause]

Hi again,

Regarding the degree of GPI, we do allow as high as degree 9, but as you said, we rarely recommend more than 3.  Results very often get unstable and unpredictable when using high-degree polynomials.  The curves will change direction one less than their degree (ex, quadratic curves bend once, cubic curves bend twice, etc), so a 9-degree polynomial will bend 8 times, and these bends can be unpredictable and not representative of the data.

If it looks like you need a degree higher than 3, we usually recommend using Kernel Interpolation or Local Polynomial Interpolation.  It is usually better to build low-degree polynomials locally than it is to build high-degree polynomials globally.

However, these are just general recommendations.  If there's something about your data where a high-degree global polynomial works best, you can of course use it.

-Eric