Social Graph Based Recommender

This invention describes a system for recommending web-based content, like blog posts or e-commerce items, to users within online social networks. It works by profiling user interests, mapping social connections, and then spreading 'endorsement' signals for items through these connections. The strength of an endorsement depends on how a user interacted with an item, and these signals are propagated across different layers of social connections to generate personalized lists of recommended items.