DoXa is a Cap digital project (french competitiveness cluster) that aims to provide a plateform for technologies related to automatic processing of moods and opinions in multilingual corpus (French and English). It is in the direct follow-up of the works done in Infom@gic project. Processed corpus gather large volumes of data unstructured - from web 2.0 (blogs, forums, social networks etc) - as well as structured ones from customer databases.

Pertimm is in charge of the sub-project dealing with moods and opinion search in the whole data set, including a task for crawling, pre-processing and filtering these data and a task for coupling the search engine with the ontologies on moods and opinions.

Doxa is structured in 8 sub-projects coordinated by Thales and it gathers 11 partners industrial (EDF, Arisem etc.), academic (Lip6/CNRS, Limsi/CNRS, IGM etc.) and SMEs (Pertimm, ILObjets etc.).