EvoMatch: An evolutionary algorithm for inferring schematic correspondences

  • Chenjuan Guo
  • , Cornelia Hedeler
  • , Norman W. Paton
  • , Alvaro A.A. Fernandes

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

1 Scopus citations

Abstract

Schema matching provides an important foundation for both manual and semi-automatic derivation of mappings between sources. However, schema matchers typically return large numbers of potentially inconsistent matches that are neither conducive to automatic mapping generation nor readily digested by mapping developers. This paper presents a method, EvoMatch, for automatically inferring schematic correspondences, from which mappings can be generated directly. It aims to offer a more expressive characterization of the relationships between sources than matches identified by existing schema matching methods. In particular, the paper contributes: i) an evolutionary search method for inferring schematic correspondences; ii) an objective function for calculating the fitness value of a solution within the search space; and iii) an empirical evaluation assessing the effectiveness of EvoMatch for inferring schematic correspondences in comparison with well established existing techniques. In doing so, EvoMatch automatically identifies correspondences that can be used directly to bootstrap information integration systems, or to inform the manual refinement of mappings.

Original languageEnglish
Title of host publicationTransactions on Large-Scale Data- and Knowledge-Centered Systems XII
EditorsAbdelkader Hameurlain, Josef Kung, Roland Wagner
Pages1-26
Number of pages26
DOIs
StatePublished - 2013
Externally publishedYes

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8320 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Fingerprint

Dive into the research topics of 'EvoMatch: An evolutionary algorithm for inferring schematic correspondences'. Together they form a unique fingerprint.

Cite this