I am estimating some spatial models. I would like to know if it is possible to include sample weights. In other words does it existna spatial counterpart to weighted least square? I tried to google this but apparently I was not able to find anything.
Thanks for any suggestion
EDIT
Basically, l estimate the following model:
y = a + bX + dWy + u
where u = kWu + e
W is a weights matrix based on the inverse distance between regions, while a, b, d and k are parameters to be estimated. So i am assuming that there exists spatial dependence on the residuals, as well as spillover effects through the spatially lagged dependent variable.
Using the model above i would like to estimate within a spatial framework the following equation:
y = a + bX + u
which is taken from a paper I am following in my analysis. The author of that paper suggests to estimate the latter equation using a WLS, adding sample weights in the estimation.
My question:
Is there anyway I can add sample weights to the estimation of the spatial model (obviously in addition to the spatial weights?).
My search so far was unsuccessful.
I hope to have clarified my question adequately.
