ETD

• erc
• innovation
• interdisciplinarity
• text-mining
• topic-modeling

In the twenty-first century, innovations arise from problem-oriented research, whose approach is oriented to cross over traditional faculties and disciplines. This leads to assume that research is increasingly moving toward more interdisciplinary endeavours. The aim of this work is to understand the conjunctural factors that favour the development of interdisciplinarity research. To achieve this objective, rather than relying on bibliometric-based measurements, as done by recent literature, we detect interdisciplinarity by analysing the textual content of a sample of the top-class european research production. Particularly, we focus on the dataset composed of the summary, final report and publications of all research projects funded by the European Research Council (ERC), and we use the machine learning technique of Topic Modeling to extract research topics and evaluate interdisciplinarity. We first built a formal model of academic research scenarios that takes into account the contextual factors that contribute to the development of interdisciplinarity research. After linking, for each academic institution awarded with an ERC project, the interdisciplinary measurements obtained via Topic Modelling with contextual variables that captured the institutional characteristics, we are able to test a series of hypotheses on which is the role played by each factor in shaping interdisciplinarity research.