How is the L-Dopa challenge scored? My impression from the webinar is that we are to provide the missing scores in the test data and the participants would be evaluated on how accurate our scores are. However, reviewing previous threads suggest that all the same features have to be used for each task per sub-challenge, so I'm guessing that I'm missing some important information. Sorry if I've missed the info somewhere and thanks for your clarification.

Created by Jennifer Schaff jeschaff
Thank you!
Within a subchallenge, all tasks must have the same set of features. As with Subchallenge 1, no missing values are allowed.
I found the scoring code. Please confirm the answer to my question regarding feature selection. Thanks-
I see that there is some scoring code available, but I am having trouble finding it, my apologies because it's probably obvious. I'm looking for the scoring code on the top wiki page (it just says it's coming soon) and on the testing data page. Thanks-
That's how I understand it, but to clarify, each of the tasks within a subchallenge needs to have the same features, correct? Scoring the drinking a glass of water and organizing papers need to have the same input of features for the bradykinesia scores, correct? Are the participants to supply the scores then, or a dataframe/table/matrix with the features that are run through your modeling algorithm? And if the later, is that scoring code available? Thanks for your information! Jenn
You do not need to provide the same feature sets for the 3 L-dopa subchallenges. You should optimize your feature sets for each specific subchallenge.

Scoring rules for L-dopa challenge page is loading…