Order-n correction for regular languages

  • Authors:
  • Robert A. Wagner

  • Affiliations:
  • Vanderbilt Univ., Nashville, TN

  • Venue:
  • Communications of the ACM
  • Year:
  • 1974

Quantified Score

Hi-index 48.27

Visualization

Abstract

A method is presented for calculating a string B, belonging to a given regular language L, which is “nearest” (in number of edit operations) to a given input string &agr;. B is viewed as a reasonable “correction” for the possibly erroneous string &agr;, where &agr; was originally intended to be a string of L. The calculation of B by the method presented requires time proportional to |&agr;|, the number of characters in &agr;. The method should find applications in information retrieval, artificial intelligence, and spelling correction systems.