The effects of orientation in a decentralized recommender
Résumé
Recommender systems are in use in every shopping system, suggesting new items to buy, new songs to listen to, new movies to watch... However, news recommendation is often still done on an explicit subscription basis, which does not yield sufficient results, being either too broad or too narrow in its scope. WhatsUp is a new distributed news recommender using implicit subscriptions to suggest links a user may want to read. Our goal was to study the effect orientation in a decentralized push-based recommender (WhatsUp), with a particular focus on content-based recommendation. We have applied several techniques to try to improve upon the current State of the Art. One of these methods was content-based, which yielded promising results for the most complete data set. This work lays the foundations for new experiments with data more closely correlated to the situation we wanted.
Loading...