Nikhil Gupta

Hi @AJarman I discussed this with @moezali1 and @Yard1 in the meeting on 20220409 and this is the recommendation. We should add an argument to setup called `forecast_limit: Optional[List[Union[int, float]]...

This one is related to the work that @Yard1 did in https://github.com/pycaret/pycaret/pull/3170 (isolating concrete property implementation to child class). Once this is merged, we can implement this as well.

Removed the no-activity tag.

I am not looking at this. Any help on documenting this would be appreciated.

You need to pass the demand column in the data argument. Currently, you are not doing it. ```python data = data[['date','internal_product_code']] ```

Also, why do you need to pass product code. The model in pycaret time series can currently only be built one time series at a time. Global modeling is not...

You can create individual models for the time series using ray and parallelization for now. An example of how to do this is shown here with M3 and M4 datasets....

Can you try to be consistent with the use of OOP and retry by updating this to OOP (I am not sure what experiment was used to load this since...

> VAR model with friendly support for differencing & reversal of differencing (ML feature scaling like friendly interface) > Auto ARIMA like feature to support friendly evaluation Could you elaborate...

Thank you. Will appreciate that!