ML2018 - Recommendation System
Setup
At first, you'll need to download all the dependencies.
Just run pip install -r environment.yml and you'll be ready.
Run
To run, just run run.py and follow the indicated steps (the values for optimal submission are the default ones, so just hit enter).
The submission corresponding to that default (optimal) values is the submission 24642 by jvonvelte with a score of 1.025.
Structure
The raw data are in the data folder.
The training set should be named data_train.csv and the sample submission set sample_submission.csv.
The output go to the reports folder.
The notebooks folder contains the Jupyter notebooks we used to develop, test and debug our project.
The core of this project is in the src folder.
The models are in src/models and the data processing are in src/data.