I am trying to apply Fay Herriot (FH) with Spatial structure (CAR=Conditional Auto Regression) using BayesSAE Package in R (here is the link to the package: https://cran.r-project.org/web/packages/BayesSAE/BayesSAE.pdf ).

However, I am so confused to provide the data for the "prox" statement. In the documentation, the author says that "PROX" is a l * 2 matrix defining the neighbourhood matrix.

I think I know how to build a neighbourhood matrix using the poly2nb() function. However, the resulting neighbourhood matrix surely does not only contain 2 columns like what the "PROX" statement required. I've tried to google this for hours but can't find any information. Can you enlighten me, please?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.