topic Re: GWR Error 040038 with single independent variable in Spatial Statistics Questions

GWR Error 040038 with single independent variable

NaciDilekli — Wed, 22 Aug 2018 20:14:04 GMT

I am trying to run GWR using a single independent/explanatory variable (atmospheric PM10) to model the relationship with a dependent variable, and it gives me the 040038 error. According to ArcMap help this means:

"040038 : Results cannot be computed because of severe model design problems. Results cannot be computed when there is either severe global or severe local multicollinearity (redundancy among model explanatory variables)".

This confuses me since there is only one explanatory variable so there can't be "redundancy among model explanatory variables". I wonder if this may be happening due to high spatial autocorrelation within the explanatory variable. What might be the issue, and how could it be fixed?

Thanks.

Re: GWR Error 040038 with single independent variable

DanPatterson_Retired — Wed, 22 Aug 2018 20:26:59 GMT

from here

Geographically Weighted Regression (GWR)—ArcGIS Pro | ArcGIS Desktop

Your data are projected and not multipoints?

What were the results of OLS?

You should always begin regression analysis with Ordinary Least Squares (OLS) regression. First find a properly specified OLS model, then use the same explanatory variables to run GWR (excluding any "dummy" explanatory variables representing different spatial regimes).

And the paragraphs that follow this quote provide further cautions.

Re: GWR Error 040038 with single independent variable

NaciDilekli — Thu, 23 Aug 2018 21:02:34 GMT

Thanks for the response! I found the problem. It is local colinearity as I suspected. I increased the kernel size and the error went away. I am looking more into this.

Re: GWR Error 040038 with single independent variable

DanPatterson_Retired — Thu, 23 Aug 2018 21:09:33 GMT

glad it is working out Naci