AutomatedProcessImprovement / knockouts-redesign

Tool for discovering improvement opportunities in knock-out checks of Business Processes
https://comserv.cs.ut.ee/ati_thesis/datasheet.php?id=75186&year=2022
Apache License 2.0
0 stars 0 forks source link

Excessive warnings with Grid Search #2

Open linomp opened 2 years ago

linomp commented 2 years ago

Suppress excessive warnings when using area under curve metric for grid search (this metric gives much better results).

It's normal that some parameter combinations during the grid search yield ill-posed training sets, but this generates a lot of warnings.

Sklearn resets the warning filters so this is very tricky.

linomp commented 2 years ago

Manuel suggested:

Instead of sklearn try another hyper-param optimizer (hyper-opt, optuna)