MetaCartSign in to MyCiteSeer

Include Citations | Advanced Search | Help

Include Citations | Advanced Search | Help

  the MeSH

Download:
Download as a PDF | Download as a PS
by Natalia Grabar, Diam Stim/dsi, Assistance Publique, Hpitaux Paris, Pierre Zweigenbaum, Pierre Zweigenbaum
http://www.biomath.jussieu.fr/~pz/FTPapiers/./Zweigenbaum:NLPBA02SUB.ps.gz
Add To MetaCart

Abstract:

Some medical resources such as the French MeSH are written without diacritic marks, which hinders their use in natural language interfaces. We examine the issue of accenting unaccented words, and propose a method for dealing with unknown words. This method learns on a reference set of accented words the minimal unambiguous contexts of the various accented forms of a given letter. We show experimental results for letter e on the French MeSH: this method proposes full accentuations for 70 % of the words that contain this letter. Address for correspondence

Citations

116 Coping with ambiguity and unknown words through probabilistic models. Computational Linguistics – Weischedel, Meteer, et al. - 1993
13 Automatic acquisition of two-level morphological rules – Theron, Cloete - 1997
7 Automatic insertion of accents in French text – Simard - 1998
2 Thirion B, et al. CISMeF: a structured health resource guide. Methods Inf Med 2000;39(1):30--5 – SJ, JP
1 Le Beux P. ADM-INDEX: an automated system for indexing and retrieval of medical texts – Seka, Courtin - 1997