<?xml version="1.0" standalone="yes"?>
<Paper uid="C02-1078">
  <Title>A Probabilistic Method for Analyzing Japanese Anaphora Integrating Zero Pronoun Detection and Resolution</Title>
  <Section position="6" start_page="25" end_page="25" type="concl">
    <SectionTitle>
5 Conclusion
</SectionTitle>
    <Paragraph position="0"> We proposed a probabilistic model to analyze Japanese zero pronouns that refer to antecedents in the previous context. Our model consists of two probabilistic parameters corresponding to detecting zero pronouns and identifying their antecedents, respectively. The latter is decomposed into syntactic and semantic properties. To estimate those parameters efficiently, we used annotated/unannotated corpora. In addition, we formalized the certainty score to improve the accuracy. Through experiments, we showed that the use of unannotated corpora was effective to avoid the data sparseness problem and that the certainty score further improved the accuracy.</Paragraph>
    <Paragraph position="1"> Future work would include word sense disambiguation for polysemous predicate verbs to select appropriate case frames in the zero pronoun detection process.</Paragraph>
  </Section>
class="xml-element"></Paper>