<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-2211">
  <Title>Speech/Language Technology Research, ETRI</Title>
  <Section position="7" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> Korean-Chinese verb patterns are invaluable linguistic resources that cannot only be used for Korean-Chinese transfer but also for Korean analysis. In the set-up stage of the development, a paper dictionary can be used for exhaustive listing of entry words and the basic usages of the words. However, as the verb patterns made from the examples of a dictionary are often insufficient, a PBMT system suffers from the coverage problem of the verb pattern dictionary.</Paragraph>
    <Paragraph position="1"> Considering there are not so many Korean-Chinese bilingual corpus available in electronic form till now, we believe the translation-based approach, i.e. Chinese-based pattern generation approach provides us with a good alternative.</Paragraph>
    <Paragraph position="2"> The focus of our future research will be given on the pre-filtering options to prevent over-generation more effectively. Another issue will be about post-filtering technique using monolingual corpus with minimized human intervention.</Paragraph>
  </Section>
class="xml-element"></Paper>