Enhancing Academic Event Participation with Context-aware and Social Recommendations

Abstract

The plethora of talks and presentations taking place at academic conferences makes it difficult, especially for young researchers to attend the right talks or discuss with participants and potential collaborators with similar interests. Participants may not have a priori knowledge that allows them to select the right talks or informal interactions with other participants. In this paper we present the context-aware mobile recommendation services (CAMRS) based on the current context (whereabouts at the venue, popularity and activities of talks and presentations) sensed at the conference venue. Additionally, we augment the current context with the academic community context of conference participants which is inferred by using social network analysis and link prediction on large-scale co-authorship and citation networks of participants. By combining the dynamic and social context of participants, we are able to recommend talks and people that may be interesting to a particular participant. We evaluated CAMRS using data from two large digital libraries - the DBLP and CiteSeerX, and participants from two conferences - ICWL 2010 and EC-TEL 2011. The result shows that the new approach can recommend novel talks and helps participants in establishing new connections at conference venue.