FairScholar: Balancing Relevance and Diversity for Scientific Paper Recommendation

摘要

In this paper, we present FairScholar, a novel scientific paper recommendation system that aims at balancing both relevance and diversity while searching for research papers in response to keyword queries. Our system performs a vertex reinforced random-walk, a time heterogeneous random-walk on the citation graph of papers in order to factor in diversity while serving recommendations. To incorporate semantically similar items in the search results, it uses a query expansion step that finds similar keywords using community detection. An online demo of our search engine is available at http://www.cnergres.iitkgp.ac.in/FairScholar/.