This repository will contain the code that solves the Pump It Up competition on drivendata.org at http://www.drivendata.org/competitions/7/

It was authored by the Waterdragers group.

Members:

Bas van Berkel

Hans-Christiaan Braun

Erik Eppenhof

Thomas van Heyningen

Senna van Iersel

Harmen Prins

All our work can be found at two places:

In the separate maps in the master branch and in the imputation branch at https://github.com/ThomasvanHeyningen/BestMLGroup2/tree/imputation

In the python map our main classification can be found in main.py, the other files in the python map are downloaded from the source given in the files.

In Preprocessing the basic feature generation methods can be found.

In the mrmr and category grouping source code for the as mapname described functions can be found.

In External Features/MLIP2 java code can be found which was used to get the temperature features.

In the data visualization some code for nice data visualizations can be found, along with the images themselves.

In https://github.com/ThomasvanHeyningen/BestMLGroup2/tree/imputation our work on imputation and pseudolabeling can be found (in the python map)

In the data map the different generated datasets can be found

All other branches are merged into the master branch and no new material is in those