Robust version of GLS with regression weights in R?
In the following webpage, I have expressed my concern about the heteroskedasticity in my residuals using rlm with regression weights in R:
I think it is probably almost always better to include the "anti-heteroskedasticity" measures in regression, because most data have that...
So I am thinking of switching to GLS...
Where do I find a version of GLS in R that has the following two features:
- Robust version.
- Can allow my own regression weights.
Thanks a lot for your pointers in advance!