Adaptive searching in succinctly encoded binary relations and tree-structured documents

  • Authors:
  • Jérémy Barbay;Alexander Golynski;J. Ian Munro;S. Srinivasa Rao

  • Affiliations:
  • David R. Cheriton School of Computer Science, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada;David R. Cheriton School of Computer Science, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada;David R. Cheriton School of Computer Science, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada;Computational Logic and Algorithms Group, IT University of Copenhagen, 2300 Copenhagen S., Denmark

  • Venue:
  • Theoretical Computer Science
  • Year:
  • 2007

Quantified Score

Hi-index 5.23

Visualization

Abstract

The methods most heavily used by search engines to answer conjunctive queries on binary relations (such as one associating keywords with web-pages) are based on computing the intersection of postings lists stored as sorted arrays and using variants of binary search. We show that a succinct representation of the binary relation permits much better results, while using less space than traditional methods. We apply our results not only to conjunctive queries on binary relations, but also to queries on semi-structured documents such as XML documents or file-system indexes, using a variant of an adaptive algorithm used to solve conjunctive queries on binary relations.