Inspired by https://github.com/sile/randomforest
Simply add the files train.csv and evaluate.csv to the ./data folder, and run ./train.sh.
The resulting labels will be created at ./out/evaluated_labels.txt.
This is a random forest classifier, the settings that work best for this data I have found to be n_trees: 200, max_depth: 12, bagging_percentage: 50%.
It reaches a consistent ~59.2% test classification accuracy on a training/testing split of 70%.