Generic Event Recognition and Extraction (GERE)
Abstract
Generic extraction was the primary focus of this funded project. To successfully extract relationships about generic entities and events, the previously developed Named Entity-based extraction algorithm was modified and extended into a process to work on events and nominal entities within the Nuclear Smuggling data domain. A secondary focus was the identification and categorization of numeric and time expressions. An evaluation shows significant improvement alone precision and recall: measures. The framework and a preliminary algorithm for specialization of event frames and attributes was designed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2002
- Accession Number
- ADA407305
Entities
People
- Eileen E. Allen
- Elizabeth D. Liddy
Organizations
- Syracuse University