[ENH] Enhancing Name Aliases for Estimators

[Kaustbh]

Describe the feature or idea you want to propose

After checking the results for all the estimators, I noticed that MR is being used as an alias for both MultiROCKETClassifier and MultiROCKETRegressor, However, I found that there are no results for MultiROCKETRegressor.

from aeon.benchmarking.results_loaders import get_estimator_results

cls = "MR"
results = get_estimator_results(estimators=cls, datasets=None, task='regression', measure='mae')
print(results)

Output:

Describe your proposed solution

Removing MultiROCKETRegressor from the alias mapping until the results are available.

Describe alternatives you've considered, if relevant

No response

Additional context

No response

[Kaustbh]

@aeon-actions-bot assign @Kaustbh

[MatthewMiddlehurst]

Thanks for looking at this. Is this the only failure from the whole dictionary for both classification and regression?

[Kaustbh]

Thanks for looking at this. Is this the only failure from the whole dictionary for both classification and regression?

Yes

After testing the whole dictionary, I noticed:

1. there are no aliases for dtw-dba.
2. estimators GRAIL, dtw-dba ( Dynamic Barycenter Averaging) has no API webpages
3. estimators 1NN-DTW have results for both classification and regression, while 5NN-DTW, 1NN-ED, 5NN-ED have result only for regression.
4. estimators like WEASEL, HC1, HC2 have results only for classification for such estimators, should we mention classifier or regressor in the aliases values? for ex. hivevoteClassifier

[Kaustbh]

After testing the whole dictionary, I noticed:
there are no aliases for dtw-dba.
estimators GRAIL, dtw-dba ( Dynamic Barycenter Averaging) has no API webpages
estimators 1NN-DTW have results for both classification and regression, while 5NN-DTW, 1NN-ED, 5NN-ED have result only for
regression.
estimators like WEASEL, HC1, HC2 have results only for classification for such estimators, should we mention classifier or regressor
in the aliases values? for ex. hivevoteClassifier

Could you please give some feedback on whether this is correct or not?

[MatthewMiddlehurst]

Dont see anything wrong there. These arent necessarily results for the current versions in aeon and not every algorithm can do both classification and regression..

[Kaustbh]

Got it, thanks for the clarification, can we add some aliases for dtw_dba as it has no aliases?