I'm wondering if there is an actual method to do the following :
Having 2 set of Identical features which are not the same training exemple as a single training exemple.
A1,A2...A10, B1,B2....B10. => A is better/ B is better
(Supervised learning) would it be possible to train a NN pr anything else that way? And then use it to. Evaluate a single. Set of feature (once training is done)?
[link][2 comments]
