A dual index model for contextual information retrieval

  • Authors:
  • Xiangji Huang;Yan Rui Huang;Miao Wen

  • Affiliations:
  • York University, Toronto, Canada;York University, Toronto, Canada;York University, Toronto, Canada

  • Venue:
  • Proceedings of the 28th annual international ACM SIGIR conference on Research and development in information retrieval
  • Year:
  • 2005

Quantified Score

Hi-index 0.01

Visualization

Abstract

In this paper, we propose a dual index model for contextual IR. For each query, we search against both document level and passage level indexes, and use the corresponding merge function to update the weights for both documents and paragraphs by combining the results from both indexes according to the granularity information in metadata. Experiments on 2004 TREC data show that a significant improvement can be made by using the dual index model.