question about final fit in MGWR class?

[ljwolf]

https://github.com/pysal/gwr_private/blob/2e666645a8fd0acf4eeec7ee18844db9f4ba484c/mgwr/gwr.py#L1057

Why do we do this second pass on the fit after already having the final XB values from the backfitting algorithim in Sel_BW?

I thought that the last iteration of the backfit is the "converged" values of each smooth?

[TaylorOshan]

Hmm, I'm not sure if I copied that over from @weikang9009 original version or if I perhaps did that when trying to fit the original version in to this GWR framework and somehow spaced out on the fact that XB was the final params. If XB.sum(axis=1) gives you the proper predicted values of y, than this can/should be massively simplified. Makes sense to me.

Maybe I started to prepare for the computation of some other diagnostics, but never followed through. Also seems I've unnecessarily reassigned the err and then subsequently done nothing with it.

[ljwolf]

I'll keep playing around with it, but one thing I do know is that the row-sum of the final XB from Sel_BW and the new predy are not the same.

Also, neither the final XB nor X * params provides us with a hat matrix that I can figure out using the substitution strategy. I'll post some of the explorations later my time.

[TaylorOshan]

Final fit is now carried over from model estimation