Latent Networks Fusion based Model for Event Recommendation in Offline Ephemeral Social Networks.

摘要: With the growing amount of mobile social media, offline ephemeral networks (OffESNs) are receiving more and attentions. Offline created ad-hoc at a specific location for purpose lasting short period time, relying on media such as Radio Frequency Identification (RFID) Bluetooth devices. The primary people in OffESNs is to acquire share information via attending prescheduled events. Event Recommendation over this kind can facilitate attendees selecting events organizers making resource planning. However, because lack users preference rating information, well explicit relations, both based traditional recommendation methods social-trust no longer work recommend OffESNs. To address challenges how derive latent preferences relations fuse unified model, we first construct two heterogeneous interaction networks, an event participation network physical proximity network. Then, use them including like-minded peers, co-attendees friends. Finally, propose LNF (Latent Networks Fusion) model under pairwise factor graph infer attendance probabilities recommendation. Experiments RFID-based real conference dataset have demonstrated effectiveness proposed compared with typical solutions.