Event Recommendation in Impromptu Event-based Social Networks

Abstract

In recent years, event-based Social Network (EBSN), which incorporates online and offline social
interactions, has experienced rapid growth and attracted a lot of research. In this paper, we propose
Impromptu Event-Based Social Network (IEBSN) as an EBSN variant, where users setup impromptu
events, get involved in event-related topics, and participate in nearby events. Impromptu events are
not supposed to be well-planned and well-organized; creation and participation of events are more
casual. Online interactions in IEBSN are not dependent on group relationship but event-based topic
recommendation, so the user’s privacy such as interest and preference is preserved from each other. A
two-level event recommendation approach is presented. Firstly, it applies the topic model to retrieve
topics from the content descriptions of event, and then uses user-topic matrix to conduct topic recommendation.
Secondly, it composes user preference, event context, organizer influence, and user’s
dynamic responses to event recommendation.