search

facebook / prophet

Tool for producing high quality forecasts for time series data that has multiple seasonality with linear or non-linear growth.
https://facebook.github.io/prophet
MIT License
18.07k stars 4.49k forks source link

how to predict thousands or millions of related time series #953

Closed tangyong closed 5 years ago

tangyong commented 5 years ago

From "DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networks" as follows:

"In recent years, a new type of forecasting problem has become increasingly important in many applications.Instead of needing to predict individual or a small number of time series, one is faced with forecasting thousands or millions of related time series."

So, I want to ask how to handle the above case using prophet?

Thanks!

bletham commented 5 years ago

There isn't currently any functionality for multivariate forecasting like this (#49 is the open issue).

I do have a comment in that issue (here: https://github.com/facebook/prophet/issues/49#issuecomment-430383681) about how if you are trying to forecast a bunch of demographic subgroups from some total then there is a way to do that using extra regressors, but that's the closest functionality there is for predicting related time series.

tangyong commented 5 years ago

Thanks reply! I will close this issue and come back #49