Skip to content

[ENH] Implement STL series to collection transformer #2851

Open
@TonyBagnall

Description

@TonyBagnall

Describe the feature or idea you want to propose

the Seasonal and Trend decomposition using Loess [1,2] is a transform that generates a collection of series from a single series. This has applications in bagging forecasting, but also could be useful for class imbalance problems.

[1] Bergmeir, C., Hyndman, R. J., & Benítez, J. M. (2016). Bagging exponential
smoothing methods using STL decomposition and Box-Cox transformation.
International Journal of Forecasting, 32(2), 303-312
[2] Hyndman, R.J., & Athanasopoulos, G. (2021) Forecasting: principles and
practice, 3rd edition, OTexts: Melbourne, Australia. OTexts.com/fpp3,
Chapter 12.5. Accessed on February 13th 2022.

Describe your proposed solution

It is implemented in statsmodels, and this seems to be what most people wrap.
https://www.statsmodels.org/dev/generated/statsmodels.tsa.seasonal.STL.html

I think we should have a bespoke, efficient, solution.

Describe alternatives you've considered, if relevant

No response

Additional context

No response

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature, improvement request or other non-bug code enhancementforecastingForecasting packageimplementing algorithmsImplementing new algorithms/estimatorstransformationsTransformations package

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions