Network Topology to Predict Bibliometrics Indices: A Case Study

Co-authorship networks have been widely studied in recent years, but today new techniques and increasing computational power permit performing novel analysis and evaluate larger datasets. One of the emerging topic is the investigation of the reasons that determine the success of some people among the others. Researchers and academic community are of interest because the metric to evaluate their performance, although widely debated, are consolidated and based on bibliometrics indices, that are quantifiable. Moreover, the paradigm of complex networks added another perspective that, often, allows discovering hidden behaviors. This paper proposes an analysis of four large datasets related to Italian academic working for public institutions, and grouped by law in academic disciplines, using network analysis tools in order to compare their structure and characteristics highlighting, if any, similarities and difference. Moreover, applying a machine learning approach, the authors try to predict some bibliometric indices using network topology.