Code for the Expedia competition on Kaggle

A simple machine learning algorithm for the Expedia competition on Kaggle.

It is designed to run with low memory (<4GB).

Content:

elaborate_features is a notebook for some prelimary analisys of the data, and generation of new features.
counter_and_leak analizes creates counters of the most common hotel_cluster with respect to the other features, then weights them to generate a predictor. It then generates a leaked solution (due to a leak in the data, since in some cases few features determine the true prediction) and updates the generated one accordingly.
the other notebooks generate a .py file to run outside the notebook to improve performance.

Name		Name	Last commit message	Last commit date
Latest commit History 7 Commits
README.md		README.md
compute_leak_file.ipynb		compute_leak_file.ipynb
counter_and_leak.ipynb		counter_and_leak.ipynb
elaborate_features.ipynb		elaborate_features.ipynb
gen_leak_solution.py		gen_leak_solution.py
leaked.ipynb		leaked.ipynb
update_submission.ipynb		update_submission.ipynb
update_submission.py		update_submission.py