text classification challenge
This is a cascading Random Forest classification model for the Tradeshift Kaggle competition. Contest details can be found here:
http://www.kaggle.com/c/tradeshift-text-classification
This code is mostly NOT my own. It was shared by a Dmitry Dryomov at http://www.kaggle.com/c/tradeshift-text-classification/forums/t/10629/benchmark-with-sklearn. All tuning was done my me however.
For the competition, I ran this a 16 core AWS instance, but it ran out of memory running the final tree. I estimate from my CV scores that this may have placed in the top 30.