Kevin Sheppard
Kevin Sheppard
You can use [threadpoolctl](https://github.com/joblib/threadpoolctl) to set the threads programmatically.
@jmakov It won't be statsmoodels - we don't touch these types of parameters.
You need to install the master copy. If you have a compiler on your system you can use ``` pip install git+https://github.com/statsmodels/statsmodels.git ``` If not, you can install a recent...
Link to the nightly is here. https://github.com/statsmodels/statsmodels/issues/6454#issuecomment-743342861
We should set Upper limits rather than pinning, e. g. numpy>=1.18.0,
We mostly work with newer versions and test against future versions. Capping can cause a lot of problems with end user setups who want to say have the latest pandas...
Should probably be `@` not `*`. Overall, I'm not sure about this in general since YW isn't identified when the matrix is singular. Might be better to raise an explicit...
You have the wrong `ARIMAResults`. You have https://www.statsmodels.org/stable/generated/statsmodels.tsa.arima.model.ARIMAResults.html#statsmodels.tsa.arima.model.ARIMAResults which does not support this function. You are using the new ARIMA, `statsmodels.tsa.arima.model.ARIMA`. The old `ARIMAResults` class and the old `ARIMA` are...
Probably should have something.
For now you can follow the example here: https://www.statsmodels.org/dev/examples/notebooks/generated/statespace_sarimax_stata.html#ARIMA-Postestimation:-Example-1---Dynamic-Forecasting