Abstract
A Natural Language Processing based Information Retrieval System that was one of the original systems developed in Phase I of TIPSTER, was the basis of research in TIPSTER III the goal of which was to add two extended capabilities to the core system. Following a description of the multiple levels of linguistic processing that were developed for the original DR-EINK System, details are provided on research into query-specific data fusion and query-specific cross-document summarization. Experimental results show that there is potential for improving retrieval through query-specific fusion and that analysts found the Detailed Multiple Document Summary to be extremely useful for almost every query, while the Thumbnail sketch was useful in approximately 50% of the queries.
Original language | English (US) |
---|---|
Pages (from-to) | 291-311 |
Number of pages | 21 |
Journal | Information Retrieval |
Volume | 3 |
Issue number | 4 |
DOIs | |
State | Published - 2000 |
Keywords
- Automatic summarization
- Data fusion
- Information retrieval
- Natural language processing
ASJC Scopus subject areas
- Information Systems
- Library and Information Sciences