Nuala A. Bennett, Qin He, Conrad Chang, Bruce R. Schatz

Digital Library Initiative (DLI) Project
CANIS : Community Systems Laboratory
University of Illinois at Urbana-Champaign
704 S. Sixth Street, Champaign, IL 61820
E-mail: {nabennet, hqin, t-chang2, schatz}
http://www.canis.uiuc.edu

This paper describes the concept extraction for the Interspace Research Project. A comparison was undertaken of four parsers for noun phrase extraction - FastNPE, NPtool, Chopper, and AZ Phraser. FastNPE was found to be the fastest of the parsers, and NPtool the most correct in extracting noun phrases. Both were subsequently implemented into the Concept Extractor module of the Interspace Prototype, which is described in detail. Future work on the Concept Extractor will include image concept extraction and this is described in the final section.