[R-sig-Geo] Imputation of missing spatial areal data

[Roger Bivand]

On Mon, 28 Oct 2019, Amitha Puranik wrote:

> Hello everyone,
>
> I would like to know whether it is possible to use the spatial 
> autoregressive model to impute missing values in aggregate data? If the 
> OLS model is replaced with SAR model in regression imputation, would it 
> lead to better estimates for missing values in a spatial data? Any 
> opinion/ suggestion is appreciated.

Please see the article referenced in the help page for 
spatialreg::predict.sarlm():

Michel Goulard, Thibault Laurent & Christine Thomas-Agnan, 2017 About 
predictions in spatial autoregressive models: optimal and almost optimal 
strategies, Spatial Economic Analysis Volume 12, Issue 2-3, 304-325

The differences in the spatial error model would be through any 
differences in covariate coefficient values, but if the differences are 
large, the Hausman test for misspecification would fail. Your post nudged 
me to raise an issue on spatialreg about SLX prediction, which very likely 
also makes sense, and to check predictions where Durbin=TRUE more 
generally.

Roger

>
> Thanks in advance.
>
> Amitha Puranik.
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

-- 
Roger Bivand
Department of Economics, Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; e-mail: Roger.Bivand using nhh.no
https://orcid.org/0000-0003-2392-6140
https://scholar.google.no/citations?user=AWeghB0AAAAJ&hl=en