[ENHANCEMENT] Make fitting with exogenous variables more relaible

[carlomazzaferro]

Is your proposed enhancement related to a problem? Please describe. Fitting with exogenous variables has no testing of any kind, and is likely broken for some of the models

Describe the solution you'd like Test cases for fitting with exogenous variables for each of the models

Describe alternatives you've considered N/A.

Additional context N/A

[vtoliveira]

I can work on that. For my undergraduate dissertation, I will use M5 data for Walmart sales up to department level (item is too much and involves intermittent demand that is another problem) and I will compare usage of classical models and ML ones with and without exogenous variables.

[rohan-gt]

Any update on this? I was looking for a way to use external regressors in hierarchical Prophet models

[carlomazzaferro]

@rohan-gt I'll try getting to this in the next week or 2

[wilfreddesert]

Hey @carlomazzaferro Are there any updates? I am currently working on a problem where having exogenous variables would really help a lot. Are you planning to fix that in the near future? If not, probably I could work on that too and make a PR.

[carlomazzaferro]

Help is definitely wanted. If you want to take a stab, please go ahead. I'm a bit backlogged but very much willing to coordinate with you

As you can tell, my initial timeline got sidetracked. Let me know if you plan on working on this to avoid double work.

@wilfreddesert

[Downfor-u]

I can provide you with a test case based on Kaggle's store item demand challenge. You can grab the data here. The objective, complete code and results so far can be found in this SO question.

In this example, we only have 4 nodes (2 stores and 2 items) for which I end up with either negative values or off the scale ones.

[jd12006]

Any updates?