<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-1710">
  <Title>Modeling of Long Distance Context Dependency in Chinese</Title>
  <Section position="5" start_page="21" end_page="21" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> This paper proposes a new MI-Ngram modeling approach to capture the context dependency over both a short distance and a long distance. This is done by incorporating long distance dependent word pairs into traditional ngram model by using the concept of mutual information. It is found that MI-Ngram modeling has much better performance than ngram modeling.</Paragraph>
    <Paragraph position="1"> Future works include the explorations of the new MI-Trigram modeling approach in other applications, such as Mandarin speech recognition and PINYIN to Chinese character conversion.</Paragraph>
  </Section>
class="xml-element"></Paper>