Time Series (create_model) - Initial PR

[ngupta23]

Related Issuse or bug

• Initial PR for create_models for time series (Work in Progress)

Describe the changes you've made

• Initial PR for create_models for time series

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Code style update (formatting, local variables)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce.

Describe if there is any unusual behaviour of your code(Write NA if there isn't)

A clear and concise description of it.

Checklist:

• My code follows the style guidelines of this project.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have made corresponding changes to the documentation.
• My changes generate no new warnings.
• I have added tests that prove my fix is effective or that my feature works.
• New and existing unit tests pass locally with my changes.
• Any dependent changes have been merged and published in downstream modules.

Screenshots

Original	Updated
original screenshot	updated screenshot

[Yard1]

I have merged time_series into this branch to take care of the X,y ordering issue using a new internal pipeline.

Just one nitpick - instead of doing isinstance() on self, compare self._ml_usecase to eg. MLUsecase.TIME_SERIES. That way imports handle better and the code is clearer.

[ngupta23]

Updates:

Splitters

• Only supporting SlidingWindowSplitter and ExpandingWindowSplitter right now.
• Not supporting timeseries splitter for now. might add back later

fh

• If integer, then forecasts that many time points ahead
  • e.g. fh = 12 will forecast from [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12] points ahead
• If list, then forecasts only those points (this needs to be tested further)
  • e.g. fh = [1, 4, 7] will forecast from only 3 points that are 1 time point, 4 time points and 7 time points ahead
• If numpy.array, then similar to list

Metrics

• Reduced metrics to only 2 for now. Needed to do this so it could work with custom cross_validate_ts function.
• Need to fix this in the long run.
• Will add more later

TBD

• add unit tests for create_model. Will add in a separate PR.
• Let's get this merged so other development can happen in parallel. I will add the unit tests right away in parallel.

[TremaMiguel]

Great work @ngupta23 just added a pair of comments.