Comparison of using passages and documents for blind relevance feedback in information retrieval

  • Authors:
  • Zhenmei Gu;Ming Luo

  • Affiliations:
  • University of Waterloo, Waterloo, Ontario, Canada;Virginia Tech, Blacksburg, VA

  • Venue:
  • Proceedings of the 27th annual international ACM SIGIR conference on Research and development in information retrieval
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper compares document blind feedback and passage blind feedback in Information Retrieval (IR), based on the work during the NRRC 2003 Reliable Information Access Summer workshop. The analysis of our experimental results shows overall consistency on the performance impact of using passages and documents for blind feedback. However, it is observed that the behavior of passage blind feedback, compared to document blind feedback, is both system dependent and topic dependent. The relationships between the performance impact of passage blind feedback and the number of feedback terms and the topic's average relevant document length, respectively, are examined to illustrate these dependencies.