Detecting Research Fronts Using Weighted Citation Networks

Abstract

In this paper, we investigate the performance of types of weighted citation network for detecting emerging research fronts by a comparative study. We employ some measures for evaluating the research fronts to weighted citation networks. For instance, average publication years and similarities of keywords are effective measures to detect research fronts. By introducing these measures as weights of citation networks to the citation network, we can detect research fronts and promising fields compared with the non-weighted citation networks. We perform a comparative study to investigate the performance of type of weighted citation networks for detecting emerging research field. Especially, we evaluate the performance of each type of weighted citation networks in detecting a research front by using the following measures of papers in the cluster: visibility, measured by normalized cluster size, speed, topological relevance, and density.