A knowledge-based search engine powered by Wikipedia

From Brede Wiki
Jump to: navigation, search
Conference paper (help)
A knowledge-based search engine powered by Wikipedia
Authors: David N. Milne, Ian H. Witten, David M. Nichols
Citation: CIKM '07 Proceedings of the sixteenth ACM conference on Conference on Information and Knowledge Management  : 445-454. 2007 November
Editors:
Publisher: Association for Computing Machinery
Meeting: Define meeting
Database(s): Google Scholar cites
DOI: 10.1145/1321440.1321504.
Link(s): http://researchcommons.waikato.ac.nz/bitstream/handle/10289/5379/milne-CIKM07.pdf
Search
Web: DuckDuckGo Bing Google Yahoo!Google PDF
Article: Google Scholar PubMed
Restricted: DTU Digital Library
Services
Format: BibTeX

A knowledge-based search engine powered by Wikipedia describes the information retrieval system, Koru, with query facility including a thesaurus based on text extracted from Wikipedia.

The thesaurus approach is also described in Extracting corpus specific knowledge bases from Wikipedia.

The thesaurus is used for automated query expansion

The researchers evaluated the performance of their Koru system by comparison with another system.

Contents

[edit] Data

  • English Wikipedia used for the thesaurus.
  • 2005 TREC HARD used for information retrieval evaluation
  • AQUAINT text corpus

[edit] Subjects

12 subjects was used for the evaluation of the system, all experience knowledge seekers. Each participant performed 10 tasks.

Experimenters observed the subjects both directly and with video. "Data was also collected from questionnaires and system log".

[edit] Results

The researchers find that the subject performs with a better recall on the Koru system and with no drop in precision.

[edit] Related papers

  1. Extracting corpus specific knowledge bases from Wikipedia
Personal tools