sktime/sktime

[ENH] Sample weights for forecasters during fit

Open

#6810 aperta il 22 lug 2024

Vedi su GitHub
 (9 commenti) (0 reazioni) (0 assegnatari)Python (1192 fork)batch import
enhancementfeature requestgood first issuemodule:forecasting

Metriche repository

Star
 (7162 star)
Metriche merge PR
 (Merge medio 26g 10h) (86 PR mergiate in 30 g)

Descrizione

Is your feature request related to a problem? Please describe. Sample weights are relevant in machine learning to describe the importance of a sample to the model. For example, regression based forecasters could support this. Currently, there is no way to push those weights to a forecaster.

Describe the solution you'd like A sample_weight argument for the fit (/fit_predict) method.

Describe alternatives you've considered No alternative available.

Additional context Some models might not support sample weights. In this case give a clear documentation description that it cannot and drop the argument during fit.

Guida contributor