Re:Is Facebook's “Prophet” the time-series Messiah or just a...
by fighterpilot on Thursday July 01, 2021 @03:02AM
(#27696963)
How well does it work with irregularly spaced time series?
Re:Is Facebook's “Prophet” the time-series Messiah or just a...
by lordgrenville on Thursday July 01, 2021 @03:44AM
(#27697145)
Just as well, in my experience, since it's curve-fitting and not autoregressive.
Re:Is Facebook's “Prophet” the time-series Messiah or just a...
by fighterpilot on Thursday July 01, 2021 @04:09AM
(#27697232)
Does it apply any kind of persistence/memory on the instantaneous exogenous variables you feed it? E.g. if you feed it the exogenous variable of "temperature right now", is it able to create a new exogenous feature "average temperature over the last three time steps"?
Re:Is Facebook's “Prophet” the time-series Messiah or just a...
by em500 on Thursday July 01, 2021 @04:22AM
(#27697284)
No, you have to handcraft all (transformations) of exogenous features. But since it's really all linear regression, that's usually reasonably straightforward.
Re:Is Facebook's “Prophet” the time-series Messiah or just a...
by microprediction on Thursday July 01, 2021 @02:38PM
(#27702753)
Do you have any city data that could be published on an hourly basis? We'd soon see how prophet does against other approaches. Just like this example:
https://www.microprediction.org/stream_dashboard.html?stream...
In terms of performance, it has been the best for a few of our forecasts, compared to GRUs, LSTM, ARIMA and SARIMA. When it wasn't the best, it wasn't too far from the best model. But, to be fair, our forecast are of quite stable data, so most models do well.
However, I would say that the key strength of Prophet is how easy it is. You can produce results really fast, you can throw data with missing range, holidays, and it has interpretability components out of the box. It depends on what do you need, but for most of our tasks, we and our stakeholders are more than happy to sacrifice a bit of performance for this features.