Can't use K-nearest weights with spatial regression anymore?

[neot...@gmail.com]

In a previous version of GeoDa for Windows on my PC, I was able to run spatial lag regression analyses using a k-nearest neighbor weights file. However, I got a new PC and just downloaded the most recent version of Geoda. Now, when I try to run the same regression analysis with that weight file, I get the following error message: "Only symmetric weights are supported for this operation. Please choose a symmetric weights file. You can still choose Classic regression for non-symmetric weights." What happened to the ability to do this in GeoDa? Is it possible to re-download previous version of GeoDa to re-run the analysis? I only see an XP version, which is otherwise too dated. Thatnks for your time.

[Julia Koschinsky]

fyi: https://geodacenter.asu.edu/node/404#crash

You can run spatial regressions with KNN weights (with non-ML estimators) in GeoDaSpace:

https://geodacenter.asu.edu/software/downloads/geodaspace

Julia

On 11 July 2014 23:07, <neot...@gmail.com> wrote:

In a previous version of GeoDa for Windows on my PC, I was able to run spatial lag regression analyses using a k-nearest neighbor weights file. However, I got a new PC and just downloaded the most recent version of Geoda. Now, when I try to run the same regression analysis with that weight file, I get the following error message: "Only symmetric weights are supported for this operation. Please choose a symmetric weights file. You can still choose Classic regression for non-symmetric weights." What happened to the ability to do this in GeoDa? Is it possible to re-download previous version of GeoDa to re-run the analysis? I only see an XP version, which is otherwise too dated. Thatnks for your time.

--
You received this message because you are subscribed to the Google Groups "Openspace List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openspace-lis...@googlegroups.com.
To post to this group, send email to openspa...@googlegroups.com.
Visit this group at http://groups.google.com/group/openspace-list.
For more options, visit https://groups.google.com/d/optout.

--

************************
Julia Koschinsky, Ph.D.
Research Director
Associate Research Professor
Arizona State University
School of Geographical Sciences and Urban Planning
GeoDa Center for Geospatial Analysis and Computation
julia.ko...@asu.edu

http://geodacenter.asu.edu
http://www.facebook.com/geodacenter
http://twitter.com/GeoDaCenter

[neot...@gmail.com]

Wow, thank you for that information. That's crazy: I even took a spatial regression course once and the professor didn't seem to know about this problem. I ran the regression on a small N and it didn't crash, but apparently the results were incorrect.

It's odd that this problem is over 10 years old now and GeoDa still hasn't been updated to address it. I guess I'll use GeoDaSpace in the meantime or try to convert the k-nearest weights to symmetric ones using R.

[Corey Sparks]

Why not just use R for everything? It will run these models using knn. Look at the spdep library, lagsarlm() and errorsarlm() functions

Corey Sparks

Assistant Professor

Department of Demography

College of Public Policy

501 West Cesar E Chavez Blvd

Monterrey Building 2.270C

San Antonio, TX 78207

corey.sparks 'at' utsa.edu

210 458 3166