<?xml version="1.0" standalone="yes"?>
<Paper uid="A92-1044">
  <Title>SEISD: An environment for extraction of Semantic Information from on-line dictionaries</Title>
  <Section position="4" start_page="253" end_page="253" type="intro">
    <SectionTitle>
4 Evaluation.
</SectionTitle>
    <Paragraph position="0"> The environment has been used to extract semantic information from the Vox dictionary. Vox is a monolingual Spanish dictionary containing about 90.000 entries (around 150.000 senses). We have concentrated on narrow but significative domains, including both noun (&amp;quot;substance&amp;quot;, &amp;quot;food&amp;quot;, &amp;quot;drink&amp;quot;, &amp;quot;person&amp;quot;, &amp;quot;place&amp;quot; and &amp;quot;instrument&amp;quot;), involving around 3000 senses, and verb (&amp;quot;movement&amp;quot;, &amp;quot;ingestion&amp;quot; and &amp;quot;cooking&amp;quot;), involving around 300 senses, taxonomies.</Paragraph>
    <Paragraph position="1"> An initial set of heuristics has been built mainly for dealing with sense disambiguation tasks. Different taxonomies have been constructed using this environment.</Paragraph>
    <Paragraph position="2"> The required linguistic knowledge sources (FPar grammars, Seg-Word rules, conversion rules) have been developped concurrently with the taxonomy building</Paragraph>
  </Section>
class="xml-element"></Paper>