philipperemy / n-beats

Keras/Pytorch implementation of N-BEATS: Neural basis expansion analysis for interpretable time series forecasting.
MIT License
861 stars 165 forks source link

Adding regressor or exogenous variable(s) #4

Closed riyadparvez closed 4 years ago

riyadparvez commented 5 years ago

Is it possible to add features that would also take regressor or exogenous variables?

Stats-model version of adding exogenous variable: https://www.statsmodels.org/dev/generated/statsmodels.tsa.statespace.sarimax.SARIMAX.html

Facebook Prophet adding regressor: https://facebook.github.io/prophet/docs/seasonality,_holiday_effects,_and_regressors.html#additional-regressors

andrewczgithub commented 5 years ago

Hi @philipperemy ! Love this work! Is the above a possibility, I know the original paper was for uni-variate series. Best, Andrew

philipperemy commented 5 years ago

Thanks guys! Yeah I agree it would be a great addition to N-beats :)

I don't have time at the moment but will put in mid term backlog!

andrewczgithub commented 5 years ago

Cheers @philipperemy ! much appreciated! Best, Andrew

philipperemy commented 4 years ago

@andrewczgithub There's an effort to support it here: https://github.com/philipperemy/n-beats/pull/12

philipperemy commented 4 years ago

cc @riyadparvez

andrewczgithub commented 4 years ago

Hi @philipperemy looking very cool!! Will this be merged? Cheers, Andrew

philipperemy commented 4 years ago

@andrewczgithub it's now merged: https://github.com/philipperemy/n-beats/commit/f89f81f4cd60c206e0862d798b83d83ac849220e Looking forward to hearing your feedbacks!