DIM_STCN

Doxavisu combines methods inspired by corpus linguistics, textual semantics, but also techniques of artificial learning and natural language processing: our methodology, already tested in the prototype # Idéo2017 is based on the description and the predictive classification of the textual and digital description of each subject analyzed, to develop a study of the argumentative and discursive specificities of the content (tweets or videos). The project combines the analysis of text and video content, and aims to provide analysis and data visualization interfaces. The goal is to give users accessible and understandable results based on technologies that are both innovative and rigorous. The project has already enabled the deployment of a prototype, presented at the Innovatives SHS CNRS 2019 event (https://25images.msh-lse.fr/innovatives-shs_2019/video/projet-cicero/fr), and aims to be implemented in other topics. This collaboration between linguists and computer scientists seems necessary to us in order to apprehend in an original and optimal way the new materials provided as digital humanities which remain to be interpreted.