My model for the See Click Predict Fix Kaggle competition. Read the postmortem for a more detailed explanation of how it works.

1. Install dependencies pip install -r requirements.txt
2. Put data in data/ directory
3. Run model python estimator.py
4. Upload submission.txt

• 147/533 on the public leaderboard
• 151/533 on the private leaderboard

Puts me in the top 70%, which is not that great, but on the bright side, my public and private leaderboard scores being very close means I didn't horribly overfit during the competition.