To run the experiment, first:
Install required Python 3 packages:
pip3 install -r requirements.txt
Download data (todo: download nltk stopwords)
wget -c "https://s3.amazonaws.com/dl4j-distribution/GoogleNews-vectors-negative300.bin.gz"
wget -c "http://deepyeti.ucsd.edu/jianmo/amazon/categoryFilesSmall/AMAZON_FASHION_5.json.gz"
python -m spacy download en_core_web_md
Perform feature extraction (this creates FAReviews_reviews.csv and FAReviews_prods.pkl)
python3 compute_scores.py
Create the matrix with distance metrics: (FAReviews_5_prods_mc.pkl)
python3 create_graph.py