[T112440833] Add Neural Prophet to KATS meta-learner

facebookresearch / Kats

Kats, a kit to analyze time series data, a lightweight, easy-to-use, generalizable, and extendable framework to perform time series analysis, from understanding the key statistics and characteristics, detecting change points and anomalies, to forecasting future trends.

MIT License

4.88k stars 534 forks source link

[T112440833] Add Neural Prophet to KATS meta-learner #248

Closed lisjin closed 2 years ago

lisjin commented 2 years ago

This PR adds Neural Prophet time-series forecasting model to the KATS metalearner.

Unfortunately, formatting the code in bb23d048fdb4c0e250d60d5e2168b39eab695df6 modified many of the files. The majority of NeuralProphet-related additions can be found in d976203a5976329fc11e1da4a928ad4bd0145480, while test cases are in 7d0ce7f364ee56185fe1e7a03aa87418532ec050.

facebook-github-bot commented 2 years ago

@lisjin has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

lisjin commented 2 years ago

@ourownstory @iamxiaodong Could you please review this PR and let me know if there are issues?

One question I have is about how to generate the true forecast outputs in kats/tests/models/test_neuralprophet_model.py. At the moment, test_forecast is just skipped.

ourownstory commented 2 years ago

@lisjin Please feel free to message me on the NeuralProphet Slack if you have any questions that I could help with!

facebook-github-bot commented 2 years ago

@lisjin has updated the pull request. You must reimport the pull request before landing.

lisjin commented 2 years ago

Closing this PR and migrating it to an internal one instead.