Efficiently enumerating results of keyword search over data graphs

  • Authors:

  • Benny Kimelfeld;Yehoshua Sagiv

  • Affiliations:

  • The Selim and Rachel Benin School of Engineering and Computer Science, The Hebrew University of Jerusalem, Edmond J. Safra Campus, Jerusalem 91904, Israel;The Selim and Rachel Benin School of Engineering and Computer Science, The Hebrew University of Jerusalem, Edmond J. Safra Campus, Jerusalem 91904, Israel

  • Venue:
  • Information Systems
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Various approaches for keyword search in different settings (e.g., relational databases and XML) actually deal with the problem of enumerating K-fragments. For a given set of keywords K, a K-fragment is a subtree T of the given data graph, such that T contains all the keywords of K and no proper subtree of T has this property. There are three types of K-fragments: directed, undirected and strong. This paper describes efficient algorithms for enumerating K-fragments. Specifically, for all three types of K-fragments, algorithms are given for enumerating all K-fragments with polynomial delay and polynomial space. It is shown how these algorithms can be enhanced to enumerate K-fragments in a heuristic order. For directed K-fragments and acyclic data graphs, an algorithm is given for enumerating with polynomial delay in the order of increasing weight (i.e., the ranked order), assuming that K is of a fixed size.