Because of their constant renewal, it is necessary to acquire fresh named entities (NEs) from recent text sources. We present a tool for the acquisition and the typing of NEs from the Web that associates a harvester and three parallel shallow parsers dedicated to specific structures (lists, enumerations, and anchors). The parsers combine lexical indices such as discourse markers with formatting instructions (HTML tags) for analyzing enumerations and associated initializers.
No takes yet. Share an insight, caveat, or question.
Jacquemin et al. (2000) studied this question.