Resources for Named Entity Recognition and Resolution in News Wires

Rosa Stern (1, 2) , Benoît Sagot (1)

1 ALPAGE - Analyse Linguistique Profonde à Grande Echelle ; Large-scale deep linguistic processing
2 Medialab AFP - Medialab AFP

In the applicative context of news wire enrichment with metadata, named entity recognition plays an important role, but requires to be followed by a resolution module that maps named entity mentions to entries in a reference database. In this paper, we describe NP, the named entity module embedded in the SXPipe shallow processing chain, that we used for extracting information from French news wires from the Agence France-Presse. We describe the construction of our reference database from freely available external resources, as well as our named entity detection, disambiguation and resolution modules. We also introduce a freely available and manually developped annotated corpus designed for the evaluation of named entity recognition and resolution tools, and provide evaluation figures for NP.

Domaines: Informatique et langage [cs.CL]

Rosa Stern, Benoît Sagot. Resources for Named Entity Recognition and Resolution in News Wires. Entity 2010 Workshop at LREC 2010, May 2010, Valletta, Malta. ⟨inria-00521240⟩