<?xml version="1.0" standalone="yes"?>
<Paper uid="P00-1037">
  <Title>An Improved Error Model for Noisy Channel Spelling Correction</Title>
  <Section position="3" start_page="1" end_page="1" type="intro">
    <SectionTitle>
2 An Improved Error Model
</SectionTitle>
    <Paragraph position="0"> Previous error models have all been based on Damerau-Levenshtein distance measures (Damerau 1964; Levenshtein 1966), where the distance between two strings is the minimum number of single character insertions, substitutions and deletions (and in some cases, character pair transpositions) necessary to derive one string from another.</Paragraph>
    <Paragraph position="1"> Improvements have been made by associating probabilities with individual edit operations.</Paragraph>
    <Paragraph position="2"> We propose a much more generic</Paragraph>
  </Section>
class="xml-element"></Paper>