Using hybrid methods and ‘core documents’ for the representation of clusters and topics: the astronomy dataset

Based on a dataset on Astronomy and Astrophysics, hybrid cluster analyses have been conducted. In order to obtain an optimum solution and to analyse possible issues resulting from the bibliometric methodologies used, we have systematically studied three models and, within these models, two scenarios each. The hybrid clustering was based on a combination of bibliographic coupling and textual similarities using the Louvain method at two resolution levels. The procedure resulted in three clearly hierarchical structures with six and thirteen, seven and thirteen and finally five and eleven clusters, respectively. These structures are analysed with the help of a concordance table. The statistics reflect a high quality of classification. The results of these three models are presented, discussed and compared with each other. For labelling and interpreting clusters, core documents representing the obtained clusters are used. Furthermore, these core documents help depict the internal structure of the complete network and the clusters. This work has been done as part of the international project ‘Measuring the Diversity of Research’ and in the framework a special workshop on the comparative analysis of algorithms for the identification of topics in science organised in Berlin in August 2014.