Auto_Arima enforce_stationarity error during fit

Question

I am following this link to create auto_arima model on my data. https://medium.com/@josemarcialportilla/using-python-and-auto-arima-to-forecast-seasonal-time-series-90877adff03c

Gitub code location is : https://github.com/Pierian-Data/AutoArima-Time-Series-Blog/blob/master/Forecasting%20a%20Time%20Series%20in%20Python.ipynb

However, during fit operation stepwise_model.fit(train) I am getting following error. Can anyone please help, I could not find and change the enforce_stationarity parameter in auto_arima constructor

ValueError: Non-stationary starting autoregressive parameters found with enforce_stationarity set to True.

Complete error below:

stepwise_model.fit(train_amt)#, freq=1)#, enforce_invertibility=False)
Traceback (most recent call last):

  File "<ipython-input-217-1bb05453004a>", line 1, in <module>
    stepwise_model.fit(train_amt)#, freq=1)#, enforce_invertibility=False)

  File "C:\Softwares\ProgramFiles\Continuum\Anaconda3\lib\site-packages\pyramid\arima\arima.py", line 343, in fit
    fit, self.arima_res_ = _fit_wrapper()

  File "C:\Softwares\ProgramFiles\Continuum\Anaconda3\lib\site-packages\pyramid\arima\arima.py", line 337, in _fit_wrapper
    **fit_args)

  File "C:\Softwares\ProgramFiles\Continuum\Anaconda3\lib\site-packages\statsmodels\tsa\statespace\mlemodel.py", line 432, in fit
    start_params = self.start_params

  File "C:\Softwares\ProgramFiles\Continuum\Anaconda3\lib\site-packages\statsmodels\tsa\statespace\sarimax.py", line 999, in start_params
    raise ValueError('Non-stationary starting autoregressive'

ValueError: Non-stationary starting autoregressive parameters found with `enforce_stationarity` set to True.

Auto_Arima enforce_stationarity error during fit

0 Answers0