Community Research and Development Information Service - CORDIS

Advanced natural language interface for multilingual text generation in healthcare

ANTHEM directly addresses two needs that are clearly identified in the healthcare community: automatic encoding and machine aided translation of medical diagnostic expressions. As such, ANTHEM is relevant for general practitioners and medical specialists using electronic medical record systems in a multilingual environment, as well as for professional encoders of diagnostic information. ANTHEM mainly serves the healthcare community by providing operational prototypes for specified needs. In addition, a number of spin-off products such as tagged corpora, lexicons, multilingual grammars for medical diagnostic sublanguage and formal representations of medical thesauri and classification systems (SNORED, ICD-10) can be used in other applications or research projects. The following results have been achieved: tagging of a Dutch/French corpus of over 200.000 medical diagnostic expressions, syntactic and semantic multilingual lexicons in Dutch, English and French relevant to the domain, a complete formalization of ICD-10 and relevant parts of SNORED, a server version of the ANTHEM language analyser, and several client demo applications. The results from ANTHEM will be useful at several levels. From a computational linguistic viewpoint, they involve substantial cross-disciplinary theoretical and operational modelling of natural language phenomena in a medical diagnostic environment.

Reported by

De Pintelaan 185
9000 Gent
See on map
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top