Indexing mixed types for approximate retrieval

  • Authors:

  • Liang Jin;Chen Li;Nick Koudas;Anthony K. H. Tung

  • Affiliations:

  • University of California, Irvine;University of California, Irvine;University of Toronto, Canada;National University of Singapore, Singapore

  • Venue:
  • VLDB '05 Proceedings of the 31st international conference on Very large data bases
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

In various applications such as data cleansing, being able to retrieve categorical or numerical attributes based on notions of approximate match (e.g., edit distance, numerical distance) is of profound importance. Commonly, approximate match predicates are specified on combinations of attributes in conjunction. Existing database techniques for approximate retrieval, however, limit their applicability to single attribute retrieval through B-trees and their variants. In this paper, we propose a methodology that utilizes known multidimensional indexing structures for the problem of approximate multi-attribute retrieval. Our method enables indexing of a collection of string and/or numeric attributes to facilitate approximate retrieval using edit distance as an approximate match predicate for strings and numeric distance for numeric attributes. The approach presented is based on representing sets of strings at higher levels of the index structure as tries suitably compressed in a way that reasoning about edit distance between a query string and a compressed trie at index nodes is still feasible. We propose and evaluate various techniques to generate the compressed trie representation and fully specify our indexing methodology. Our experimental results show the benefits of our proposal when compared with various alternate strategies for the same problem.