> The final model evaluation metrics


I want to ask at the final stage, all models will be ranked by what evaluation metrics?

Thank you so much!

Posted by: wyn2168 @ Sept. 23, 2022, 4:42 p.m.

Hi, rankings will be based on the area under the ROC curve with log-loss as the secondary metric in case of ties (see the Challenge Details), the same metrics you get for the practice cases (for illustrative purposes only since performance metrics for only 12 practice cases have large error bars).

Posted by: kdrukker @ Sept. 23, 2022, 5:04 p.m.

Thank you so much! I didn't notice that before.

Posted by: wyn2168 @ Sept. 23, 2022, 6:22 p.m.
Post in this thread