Abstract

The rating of Computer Science (CS) conferences are important as it influences how papers published at the conferences and may also be used to evaluate research. In this paper, we proposed a method, rsIT, based on a small given set of top conference (pivots) and a relatedness measure based this set as well as basic baseline methods using citation count and field rating. We experimented with a snapshot dataset from Microsoft Academic Graph together with conference data from Microsoft Academic Search. We evaluated the conference ratings from our methods with the CCF conference rating list. We showed that rsIT correlates well with CCF rating and correlates better than ratings from using a baseline ranking with citation count or field rating.