Nikhil Gupta

Related to multivariate time series - see https://github.com/pycaret/pycaret/issues/1588 as well as Time Series with exogenous variables.

Example of TSFresh with sktime: https://github.com/alan-turing-institute/sktime/blob/main/examples/feature_extraction_with_tsfresh.ipynb Also, my analysis here: https://gist.github.com/ngupta23/bdd91ea76519abfb48355c6ea5cefcd4

@goodwanghan - Can you help with this? Thanks!

Now that we have a dedicated time series module, it may also be worthwhile to have a look at that (it requires less feature engineering on your end). I converted...

Are you using the latest 3.0.0 release? It should ideally have a much lower dependency footprint than the 2.x release.

No, I don't. I just know that it should have a much smaller dependency. Also, since the NLP module was removed in 3.0.0, all the big language dictionaries that were...

Well, it was originally intended to be on the predictions only. The idea was that the input being fed would already be limited to the correct values. But the model...

Awesome. Please give it a shot. This is what I was thinking. (1) Introduce 2 new parameters in `setup` called `lower_clamp` and `upper_clamp` (defaults to `None` which implies no limits)....

> @ngupta23 I assume this would be useful in the regression module also? I think it would be. @moezali1 @Yard1 Would you like to include this functionality in the regression...

Yes, fork from `time_series` and merge back to the same branch. Thanks!