<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-2114">
  <Title>Rebuilding the Oxford Dictionary of English as a semantic network</Title>
  <Section position="10" start_page="5" end_page="5" type="concl">
    <SectionTitle>
6 Strength of association from Domain_A to
</SectionTitle>
    <Paragraph position="0"> Domain_B is determined internally to ODE, by calculating the proportion of (a) lexemes and (b) semantic sets in which both domains appear, as opposed to those in which only Domain_A appears. Strength of association is not mutual.</Paragraph>
    <Paragraph position="1"> dictionary, albeit one that takes advantage of novel search and navigation features. Additionally, ODE is directly usable in a number of non-dictionary applications. These include context-sensitive spellchecking, tagging and parsing, document categorization, and context-sensitive document glossing. Feedback from such applications is being monitored not only to critically examine and correct the source data, but also to examine the source dictionary itself: because much of the formal and classificatory data is generated algorithmically from analysis of the source editorial content of each lexical object (definition, etc.), anomalies emerging in that data can often be traced back to anomalies in the editorial content (e.g. inconsistencies in defining style). The ODE project is therefore in part an attempt to bridge the distinction between human-user dictionaries and WordNet-like associative electronic lexicons. By deriving the one from the other, it invites applications to navigate and mine the rich lexicographic content of the original dictionary by means of a new set of structured relations and frameworks.</Paragraph>
  </Section>
class="xml-element"></Paper>