Abstract
The development of science in various fields have prompted researchers to represent the scientific literature as a network whose nodes are works and citations are the links (citation network). One major problem is to be able to carry out an efficient clustering which allows grouping the articles according to their similarities. This method should be automatic and must not involve a careful reading of the selected works. In this work we have developed a new clustering method for a citation network using a definition of distance introduced by Bommarito el al. (2010). Moreover through the analysis of the adjacency matrix as a Recurrence Plot, it is possible to view the cluster evolution over time.