<?xml version="1.0" standalone="yes"?>
<Paper uid="E93-1059">
  <Title>Enhancing a large scale dictionary with a two-level system</Title>
  <Section position="2" start_page="1" end_page="1" type="abstr">
    <SectionTitle>
3 Results
</SectionTitle>
    <Paragraph position="0"> We tested this transducer on a 1,300,000 words corpus containing 58,000 different graphical forms. Our transducer analyzed 75% of these graphical forms, which is 3% more that the transducer of DELAF alone, at a speed of 1,100 words per second. Hence, more than 97% of the word occurrences of our corpus have been analyzed in the following way: algorithmisation ~ algorithme,N*iser,V*Ation,N* f*.s</Paragraph>
  </Section>
class="xml-element"></Paper>