<?xml version="1.0" standalone="yes"?>
<Paper uid="I05-2012">
  <Title>Automatic Extraction of English-Korean Translations for Constituents of Technical Terms</Title>
  <Section position="6" start_page="71" end_page="71" type="concl">
    <SectionTitle>
5 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper, we have described an alignment algorithm between English and Korean term constituents. Our alignment algorithm can handle cross alignment, n:1 alignment and 1:n alignment between term constituents. Our method shows about 94.7% precision, 93.2% recall and 6.1% alignment error rate. However, there are scopes to improve performance still further. Constraints should be relaxed in order to generalize our model and overcome errors caused by them.</Paragraph>
    <Paragraph position="1"> Our method can be applied to handle technical terms in three aspects. First, alignment results produced by our alignment algorithm help a machine translation system to consistently translate new English technical terms to Korean terms by considering domain of the technical terms. Second, alignment results between term constituents can be used for constructing term formation patterns or word formation patterns.</Paragraph>
    <Paragraph position="2"> Because relations between conceptual units can be extracted from the alignment results, we can construct concept-level term formation patterns using them. Third, the alignment results can be used as a resource for recognizing term variations. Because alignment relations acquired by our alignment model offer information about homonym, synonym and domain dependency, term variations related to certain term constituent can be recognized using them.</Paragraph>
  </Section>
class="xml-element"></Paper>