hi,
At SEA/SEW in Paris, I learned about Smirnov (2020) which provides a method to estimate rho in a lag model using regressions on y, Wy, and W^2y. Genieaux uses this interestingly to build a very fast GAM framework for SAR geo-additive models, available since last week.
This looks very interesting, since we could do something similar with pygam.
I'm sure, though, fully nonlinear spline models may be out of scope for spreg?
hi,
At SEA/SEW in Paris, I learned about Smirnov (2020) which provides a method to estimate rho in a lag model using regressions on y, Wy, and W^2y. Genieaux uses this interestingly to build a very fast GAM framework for SAR geo-additive models, available since last week.
This looks very interesting, since we could do something similar with
pygam.I'm sure, though, fully nonlinear spline models may be out of scope for spreg?