Unsupervised characterization of research institutions with task-force estimation

•The relevant task-force is established from bibliographic information.•Different types of authors in articles are defined.•The most relevant sets of authors (effective core members and collaborators) included in the task-force of an institution are identified.•Time evolution of the task-force involved in scientific production is easily established.•New indicators based on task-force estimation are defined.

This work aims at establishing the task-force involved in scientific production at the institutional or national level, globally or per area or sub-area of knowledge. In the proposed system, the estimated task-force is further divided into core (permanent members of the institution(s)) and collaborators (more mobile members), and allows normalization of scientific production.Research groups/institutions/countries of different sizes/scientific areas can, thus, be directly compared and the time evolution of these groups inspected. Results are presented for the characterization of four universities (from Portugal, Sweden and USA) in the 2008–2012 period, for the research area of Chemistry. It is shown that it is possible not only to estimate the task-force, but also to derive new, relevant indicators for the set under analysis. Aspects pertaining to collaboration fluxes are also assessed.