Fil received a $10k Yahoo! Faculty Research and Engagement award to study methodologies to infer user intents and goals based on social similarity measures. This project is in collaboration with Debora Donato of Yahoo! Labs, and Luca Aiello and Giancarlo Ruffo of the University of Torino, Italy.
Fil will present the paper Folks in folksonomies: Social link prediction from shared metadata (authored with Rossano Schifanella, Alain Barrat, Ciro Cattuto, and Ben Markines) at WSDM 2010 in New York on February 5. The paper discusses homophily, or more specifically the relationship between social connections and social tagging in folksonomies. We show that social similarity measures based on annotations can be effective predictors of friendship relationships. For the occasion, we are making our Last.fm dataset publicly available.