theusual/kaggle-seeclickfix-ensemble

Prize winning solution to the SeeClickFix contest hosted on Kaggle, developed by teammates Bryan Gregory and Miroslaw Horbal. The purpose of the contest was to train a model (as scored by RMSLE) using supervised learning that will accurately predict the views, votes, and comments that an issue posted to the www.seeclickfix.com website will receive. My teammate and I used this ensemble code base to combine our top ranked individual models to create a prize-winning solution, defeating >500 other teams and winning a prize of $1,000.

PythonBSD-3-Clause