search

alkaline-ml / pmdarima

A statistical library designed to fill the void in Python's time series analysis capabilities, including the equivalent of R's auto.arima function.
https://www.alkaline-ml.com/pmdarima
MIT License
1.57k stars 231 forks source link

Will support for exogenous variables be removed? #488

Closed paulsbrookes closed 2 years ago

paulsbrookes commented 2 years ago

Describe the question you have

I have noticed comments in the code which suggest there is a plan to remove support for exogenous variables. For example in line 779 of pmdarima.arima.arima.ARIMA.predict_in_sample I see the following:

    check_is_fitted(self, 'arima_res_')

    # Temporary shim until we remove `exogenous` support completely
    X, kwargs = pm_compat.get_X(X, **kwargs)

When is this change planned?

aaronreidsmith commented 2 years ago

Support for exogenous variables will not be removed, the argument is just being renamed to X. See the RFC for more details around why it was changed.

As far as when it will be changed, we plan to make the cutover in version 2 (which we don't have a release date for yet). Right now you will get a warning if you pass in the exogenous kwarg. If you are doing so in your code, simply change it to X and the warning will disappear and you will be compatible with future versions.

paulsbrookes commented 2 years ago

Thank you.