Abstract
This paper presents a Named Entities (NE) recognition system for the English written language, which combines the wealth of the WordNet taxonomy and the effectiveness of traditional rule-based approaches. The core of the system relies on the combination of approximately 200 language-dependent rules with a set of predicates, defined on the WordNet hierarchy, for the identification of both proper nouns and trigger words. The strengths of this approach are twofold. First, the use of a semantic network allows it to cope with the difficulty of building and maintaining extensive gazetteers. Second, considering the recent spread of WordNet-like semantic networks for languages other than English and aligned with the English version, the use of language-independent predicates offers a useful basis for achieving multilinguality.
Users
Please
log in to take part in the discussion (add own reviews or comments).