<?xml version="1.0" standalone="yes"?>
<Paper uid="C86-1067">
  <Title>A Kana-Kanji Translation System for Non-Segmented Input Sentences Based on Syntactic and Semantic Analysis</Title>
  <Section position="6" start_page="284" end_page="284" type="concl">
    <SectionTitle>
5. CONCLUSION
</SectionTitle>
    <Paragraph position="0"> A new disambiguation approach to Kana-kanji translation for non-segmented input sentences has been described. Ambiguitiy is resolved using syntactic and semantic analysis based on a case grammar. To avoid a combinatorial explosion of the ambiguity, some heuristics are introduced. Large dictionaries were also developed for the experimental system and both the limit and substantial performance of the system were evaluated.</Paragraph>
    <Paragraph position="1"> The experimental results show that an accuracy of 90.5% is obtainable using this approach, and that the accuracy can be improved to about 95% by optimizing the dictionaries. Further improvement can be achieved by introducing context analysis and plagmatic analysis.</Paragraph>
  </Section>
class="xml-element"></Paper>