A personalized people recommender system using global search approach

Abstract

The goal of people recommender system is to generate meaningful social suggestion to users. The abundant data are the key factor in fulfilling a recommendation task, but the cost of user data in a real-world system is high. In this paper, we propose a novel approach that integrates a global search result with a personalized people recommendation system. Our approach utilizes the user identity as a query keyword and processes the search results through five different customized parsers. This approach solves the cold-start issue in recommendation systems and leverages the cross-domain information in order to provide a better recommendation result. To test our approach, we embedded it into an existing conference navigator system then deployed the system at two international conferences. The survey results indicate largely positive feedback about the system's effectiveness. Our study results also shed some light on the social interactions that take place at an academic conference.