Describe the feature you want to add to this project

Currently, seasonality defaults to multiplicative if it is allowed, only switching to additive when multiplicative is not allowed. However, there is no formal test for which one should be used.

Current benchmarking results indicate that the performance of ETS in pycaret on M3 dataset is slightly less than statsforecast. This could be a result of seasonal detection.

Describe your proposed solution

Try to improve the seasonal detection with the proper seasonality type and see if it improves the results

Describe alternatives you've considered, if relevant

None

Additional context

No response