My solution ranked 8th out of 2216 on the Recruit Restaurant Visitor Forecasting Kaggle competition. The solution focuses on targeted feature engineering and LightGBM cross-validation.
- Have Python 3 installed
- Download the Kaggle data from here
- Download the weather data from here
- Run
pip install -r requirements.txt
(using a virtual environment is good practice) - Install LightGBM
- Run
jupyter notebook
- Open the
Solution.ipynb
notebook