Parameter optimization: Add Early stopping for all strategies

Simple feature request: Could you add Early stopping option for all strategies? Similar to neural networks. Stop if no improvement was found over configurable number of iterations? Or at least add the option to TPE, so that one can theoretically choose high number of iterations but stop if it’s not worth it?

1 Like

Hi @beginner,

I like this feature request and created a ticket in our internal system (AP-14859). Thanks for reporting it!



This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.