This is a submission for the Kaggle What's Cooking Challenge. It was built to serve as a simple example of applied machine learning.

See the post for a full explanation and code overview.

Quick Start

1. Download the supporting files from Kaggle
2. Update filename variables in each of the .py files as needed
3. Run parser.py to take the training set from Kaggle and reformat it for our needs
4. Run train.py to train the model and generate predictions over the Kaggle test set
5. Run kaggle.py to reformat the predictions into the CSV reppresentation that Kaggle expects