How many feature dimensions are supported for supervised learning?

User 117 | 9/29/2014, 10:07:17 AM


We are currently working on a supervised learning problem where the dimensionality of the features can range from 1440 features all the way up to 50400 dimensions.

So I was wondering if there exist some guidelines about the maximum number of features each supervised learning method can support.

Also, are there any examples on how to use Factorization Machines for binary classification through GL Create?


User 91 | 9/29/2014, 8:07:03 PM

Right now, we don't have a way to do binary classification using Factorization machines, but we have it in our roadmap and you can expect it soon.

Right now, you should have no issues with problems of the dimensions that you have mentioned. We have tested it to work on much larger feature dimension sizes. Do let us know if you have any issues with any of the models.