<?xml version="1.0" standalone="yes"?>
<Paper uid="P03-1016">
  <Title>Synonymous Collocation Extraction Using Translation Information</Title>
  <Section position="7" start_page="21" end_page="21" type="concl">
    <SectionTitle>
4 Conclusions and Future Work
</SectionTitle>
    <Paragraph position="0"> This paper proposes a novel method to automatically extract synonymous collocations by using translation information. Our contribution is that, given a large monolingual corpus and a very limited bilingual corpus, we can make full use of these resources to get an optimal compromise of precision and recall. Especially, with a small bilingual corpus, a statistical translation model is trained for the translations of synonymous collocation candidates. The translation information is used to select synonymous collocation pairs from the candidates obtained with a monolingual corpus. Experimental results indicate that our approach extracts synonymous collocations with an average precision of 74% and recall of 64%. This result significantly outperforms those of the methods that only use monolingual corpora, and that only use a bilingual corpus.</Paragraph>
    <Paragraph position="1"> Our future work will extend synonymous expressions of the collocations to words and patterns besides collocations. In addition, we are also interested in extending this method to the extraction of synonymous words so that &amp;quot;black&amp;quot; and &amp;quot;white&amp;quot;, &amp;quot;dog&amp;quot; and &amp;quot;cat&amp;quot; can be classified into different synsets.</Paragraph>
  </Section>
class="xml-element"></Paper>