GitHub - nikhil1008/social-graph-based-recosys: I tried to explore a new approach to recommender systems which can utilize information in social networks, including user preference, item's general acceptance, and influence from social friends. I could achieve on par performance with standard CF based recommender systems. On the other hand, this system doesnt suffer from cold-start data sparsity problems common to CF based systems . This project uses real-world dataset from Yelp challenge which consists of 1637138 users and 108179 businesses.

nikhil1008 / social-graph-based-recosys Public

Notifications You must be signed in to change notification settings
Fork 0
Star 0

I tried to explore a new approach to recommender systems which can utilize information in social networks, including user preference, item's general acceptance, and influence from social friends. I could achieve on par performance with standard CF based recommender systems. On the other hand, this system doesnt suffer from cold-start data sparsi…