ChadFulton

Thanks for reporting this. I think we should probably be able to support forecasting after removing data, so this is probably just about what assumptions the forecasting code makes (e.g....

Thanks for the bug report! This must have something to do with a bad value getting into one of the arrays being decomposed. I can't replicate any LU decomposition error...

(previous comment was submitted too fast) Ultimately the issue here is the behavior when the parameters imply a non-stationary model, but only barely. I would guess that the symptom of...

Thanks for checking in. I guess that the linear algebra error doesn't show up in all non-stationary cases, but I can replicate the LU decomposition error with e.g.: ```python mod...

I have not used rolling OLS extensively for anything myself, but I suppose that the appropriate residuals depend on what you plan to do with them. For example, in your...

> Re DOI, apologies. some more searching turned up the relevant text > > Harvey, A. (1990). Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press....

> statsmodels.tsa.statespace.SARIMAX on pd.Series and add one exogeneous variable - i see load on all cores despite mentioned settings. Morover, if i add not one, but two or more exogeneous...

@codeananda thank you for the PR

See https://stackoverflow.com/questions/61991259/ for evidence that non-full period indexes do not work (at least right now).

There's no way to do this automatically right now, but it is a good idea. Still, we don't have anyone actively working on VARs right now, so I don't have...