Research collaboration and topic trends in computer science: An analysis based on UCP authors

Abstract

Academic publication metadata can be used to analyze the collaboration, productivity and hot topic trends of a research community. Recently, it is shown that authors with uninterrupted and continuous presence (UCP) over a time window, though small in number (about 1%), amass the majority of significant and high-influence academic output. We adopt the UCP metric to retrieve the most active authors in the Computer Science (CS) community over different time windows in the past $50$ years, and use them to analyze collaboration, productivity and topic trends. We show that the UCP authors are representative of the overall population; the community is increasingly moving in the direction of Team Research (as opposed to Soloist or Mentor-mentee research), with increased level and degree of collaboration; and the research topics become increasingly inter-related. By focusing on the UCP authors, we can more easily visualize these trends.