Mixing Statistical and Symbolic Approaches for Chemical Names Recognition - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2008

Mixing Statistical and Symbolic Approaches for Chemical Names Recognition

Résumé

This paper investigates the problem of automatic chemical Term Recognition (TR) and proposes to tackle the problem by fusing Symbolic and statistical techniques. Unlike other solutions described in the literature, which only use complex and costly human made ruled-based matching algorithms, we show that the combination of a seven rules matching algorithm and a na¨ve Bayes classifier achieves high performances. Through experiments performed on different kind of available Organic Chemistry texts, we show that our hybrid approach is also consistent across different data sets.

Dates et versions

hal-01313224 , version 1 (09-05-2016)

Identifiants

Citer

Florian Boudin, Marc El Bèze, Juan-Manuel Torres-Moreno. Mixing Statistical and Symbolic Approaches for Chemical Names Recognition. 9th International Conference, CICLing, Feb 2008, Haifa, Israel. ⟨10.1007/978-3-540-78135-6_28⟩. ⟨hal-01313224⟩

Collections

UNIV-AVIGNON LIA
36 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More