Dear organizers,
I have a quick question: Do we need to select one of the two submissions for round 1 ranking? Or it will automatically pick the better one?
Another suggestion for this, could you also add some baseline model to the ranking (e.g. simple regression or model using only the clinical data) ? So it can tell whether our submissions have prediction power and can guide us to improve it later.
Thank you.
Jincheng
Created by Jincheng Wu Jincheng2009 Hi Mike,
Okay, I see. Sadly after some significant "improvement", my second submission scored worse than the first one...
Thank you for explaining, and yes, any baseline model would help.
Jincheng
Dear Jincheng,
We will just take the last model submitted by the team. Not prizes are awarded in leaderboard rounds in this challenge so these rankings have little overall impact in the challenge.
We have indeed discussed including baseline models in the leaderboards but removed them. I will discuss this in our next internal meeting and possibly add in baseline / ref models.
Regards,
Mike