Hello organizers, Is it possible to get more explanations for what you mean by "features"? I am guessing that there are minimum and maximum numbers of features we can provide? Technically speaking, we could provide predictions from a complex model as 1 dimensional feature for a simple model. Is that allowed?

Created by Naozumi Hiranuma nhiranuma
Dear @nhiranuma: The goal is to submit features that in aggregate are predictive of disease state. By feature we mean a numerical measurable property of the phenomenon being measured. In sub-challenge 1 we will be predicting a binary classification but there are perturbations to the disease state and information about disease severity that will be used to evaluate your features as well. As such, a single feature might work for the prediction of disease state but will very likely not generalize to the collaborative phase.

Could we get the definition for "features"? page is loading…