[Openspace] Heteroskedasticity

Vania Ceccato vania at infra.kth.se
Tue Sep 18 04:33:18 CDT 2007


My models show strong signs of heteroskedasticity even after using Lag or Error
models. Looking at my SpaceStat manual(chapter 30), Anselin suggests using
ML-Ghet model to try to solve this problem (which is not implemented in GeoDa).
I still have SpaceStat in my PC so I can test ML-Ghet model. The problem is that
I don't know how to input the variables into the model properly. I need a
heteroskedasticity variable, which I think should be categorical, but how
should this variable look like? as districts of my city, from 1 to 10 for

I am grateful for any suggestions to solve the problem of heteroskedasticity.

Thank you


Vania Ceccato
Department of Urban Planning and Environment 
Royal Institute of Technology

More information about the Openspace mailing list