r2 linear = 0.72
r2 dec.tree = 0.7
participants have to choose between equally performing or better. I find the difference small so i choose for equally performing → wrong answer
I suggest to make such questions a bit more specific and less dependant on estimation judgement.
True, we modified Q. 4 to remove the ambiguity but we forgot to change Q. 3. Sorry about that.
@lfarhi @MarieCollin Could you make the change in FUN: https://gitlab.inria.fr/learninglab/mooc-scikit-learn/mooc-scikit-learn-coordination/-/commit/21fc0279863f763b15946ed82bb0e32d9a7e61a4
The change has been done in: FIX remove ambiguity · INRIA/scikit-learn-mooc@e794518 · GitHub It will be soon made available in FUN.
Mode whisper : @glemaitre58 est-ce que je réinitialise les réponses ?
A mon avis pas besoin. Il y a peu de répondants, je peux aussi proposer dans ce fil de réinitialiser pour ceux qui le souhaitent
It is available in FUN now.
On peut laisser comme ca.
Sorry but i think the question is still ambiguous.
You propose as possible answers " (better/equally/worse) than a linear model: (~0.74/~0.72/~0.7)" . the problem here is the building of the sentence could lead to think that the number after the ‘:’ are the scores of a linear model.
It could be less ambiguous if the questions were asked as :
A tree with an optimal depth has a score of :
a) ~0.74 and is better than the linear model
b) ~0.72 and is equal to the linear model
c) ~0.7 and is worse than the linear model
Thanks for the wording. I made the change here: FIX add suggestion · INRIA/scikit-learn-mooc@68742d5 · GitHub
@lfarhi @MarieCollin could you make the change from https://gitlab.inria.fr/learninglab/mooc-scikit-learn/mooc-scikit-learn-coordination/-/commit/47e7a6bf374546f83f80abb981c92a66ebde7384
ok, it’s fixed !
