Rare Diseases Symptoms Automatic Extraction


This topic is proposed as part of the ANR Hybride project involving three research laboratories ( MoDyCo , LORIA , GREYC ) and one French research institute (INSERM) . The ANR Hybrid project, started in 2011 , combines the methods of two areas: the automatic language processing and data mining . Its objective is to develop new methods and tools for knowledge discovery in textual data. The project will be applied to the field of rare diseases to help experts' diagnosis of such diseases and hopefully discovers until then ignored links between symptoms and orphan diseases.

This project was lead by four students who shared everything from learning about previous work in medical data-mining to constructing this website and of course including analysing already known symptoms to discover how they are linguistically expressed and set up a processing to expose new symptoms in a corpus of 10,000 medical abstracts. We used scripting languages and specialized softwares.