<?xml version="1.0" standalone="yes"?>
<Paper uid="N03-3004">
  <Title>Discriminating Among Word Senses Using McQuitty's Similarity Analysis</Title>
  <Section position="10" start_page="0" end_page="0" type="concl">
    <SectionTitle>
9 Conclusions
</SectionTitle>
    <Paragraph position="0"> We have presented an unsupervised method of word sense discrimination that employs a range of surface lexical features, and relies on similarity based clustering. We have evaluated this method in an extensive experiment that shows that our method can achieve precision and recall higher than the majority sense of a word for a reasonably large number of cases. We believe that increases in the amount of training data employed in this method will yield to considerably improved results, and have outlined our plans to address this and several other issues.</Paragraph>
  </Section>
class="xml-element"></Paper>