This repository includes my scripts for the Kaggle Facebook Recruiting III - Keyword Extraction competition. My solution is based on Alex Minnaar's blog post.
- The
model
folder includes the following scripts:pre_process.py
which prepares the training and testing files,classification_model_mongodb.py
which creates association rules and creates the prediction output. - In addition, you will need to create a
csv
directory and copy the data files in there (download data csv files here).
I would appreciate any suggestions you may have regarding improving the scripts. My current best public score on Kaggle is 0.50060.