CoRank: Simultaneously Ranking Publication Venues and Researchers

Byron J. Gao, Gayathri Karupakula Jagadeesh Kumar

Abstract: Many academic and administrative decisions rely on evaluation of publication venues and individual researchers, such as for the purpose of hiring, promotion, and grant distribution. Thus it is important to compare and rank publication venues and researchers in an objective and authoritative manner. While many endeavors exist, in this paper we propose a novel algorithm CoRank based on venue-researcher interaction that computes reputation scores and rankings for publication values and researchers simultaneously. We observe that good researchers publish many papers in good venues, good publication venues feature many good researchers, and venue scores and researcher scores can be defined in terms of each other. CoRank is designed to break this circular definition in an iterative manner to compute reputation scores and rankings for both venues and researchers. We implement CoRank and perform experiments on real DBLP dataset to demonstrate its promise.