search

intive-DataScience / tbats

BATS and TBATS forecasting methods
MIT License
178 stars 19 forks source link

Editable Reproduction (Stationarity Issue) #11

Closed firmai closed 4 years ago

firmai commented 4 years ago

What could be wrong with my data it works on pmdarima. But not with TBATS.

https://colab.research.google.com/drive/17Ye4byHITNn4DhEqup_mXfWLycgjNhuf

Even after making the data stationary (differencing) it still screamed the same error.

Thanks in advance, and nice clean package!

Regards!

anqus commented 4 years ago

I think I'm having the same issue as you. Do you get the following error?

ValueError: Could not successfully fit ARIMA to input data. It is likely your data is non-stationary. Please induce stationarity or try a different range of model order params. If your data is seasonal, check the period (m) of the data.

My notebook was working when I last ran it (maybe 6 weeks ago, nothing has changed since). I've tried dropping back to TBATS 1.0.8, (1.0.9 was released 8/01/2020) but no success.

firmai commented 4 years ago

Same issue, not sure what could be causing this. Good to know it's not isolated.

cotterpl commented 4 years ago

I will have a look into this soon. In the mean-time try using it with use_arma_errors=False

lukedyer-peak commented 4 years ago

This seems to be related to the statsmodel 0.11.0 update. I went back to 0.10.2 and didn't get this bug anymore.

cotterpl commented 4 years ago

I have reviewed the issue. It lies inside pmdarima package that fails to work properly with statsmodel that @lukedyer-peak mentions. It works properly with 0.10.2. I am closing this as non TBATS related.