Insights into the French socio-ecological research network through Natural Language Processing
29 October 2021
Scientific research on SocioEcoSystem (SESs) has grown exponentially since the seventies, but because of the heterogeneity in the actors, the disciplines and the collected data, some efforts are still necessary to build a common language anda thesaurus for indexing data.
Natural Language Processing (NLP) methods were used to analyse a French corpus derived from the 5th colloquium of the French long term socio-ecological research network RZA (Réseau des Zones Atelier) which marked the network's 20th anniversary.
The authors (Ingrid Falk and Isabella Charpentier) investigated the involved vocabulary to cross reference the subjects of interest and explore how well automatically extracted topics are related to the ambitions of the RZA community in terms of inter- and trans-disciplinary environmental research.
According to this topic analysis, the RZA was found to go beyond the so-called disciplinary spheres, as it carries out trans- and inter-disciplinary studies in the fields of social and natural sciences, on different socio-ecosystems, and notably hydro-systems, urban environments and rural areas (agro-ecology).
Combined with an essential and indispensable domain expertise, NLP techniques allowed a much more structured and in depth content analysis than the simple frequency analysis, offering opportunities for the completion of the EnvThes thesaurus.
Access the article here.
