<?xml version="1.0" standalone="yes"?>
<Paper uid="C04-1050">
  <Title>Improving Japanese Zero Pronoun Resolution by Global Word Sense Disambiguation</Title>
  <Section position="6" start_page="0" end_page="0" type="evalu">
    <SectionTitle>
5 Experimental Results and
</SectionTitle>
    <Paragraph position="0"> We conducted experiments of zero pronoun resolution on two different domain corpora. One is newspaper articles of &amp;quot;Relevance-tagged corpus&amp;quot; (Kawahara et al., 2002), and the other is utterances of cooking TV programs. These cooking utterances were handled by (Shibata et al., 2003). They annotated various relations to closed captions of the cooking utterances based on the specification of the &amp;quot;Relevance-tagged corpus&amp;quot; (Kawahara et al., 2002).</Paragraph>
    <Paragraph position="1"> For newspaper domain, the antecedent preference and the classifier were trained with 1,841 sentences in the newspaper corpus, and the newspaper case frames were used. The experiment was performed on 633 sentences. For cooking domain, we used 813 sentences (5 TV programs), and conducted 5-fold cross validation using the cooking case frames.</Paragraph>
    <Paragraph position="2"> We evaluated &amp;quot;ga&amp;quot;, &amp;quot;wo&amp;quot; and &amp;quot;ni&amp;quot; cases that are the large majority of zero pronouns. The experimental results are shown in Table 2 and  culated by evaluating both detection and antecedent estimation of zero pronouns together.</Paragraph>
    <Paragraph position="3"> The baseline corresponds to our previous system without word sense disambiguation.</Paragraph>
    <Paragraph position="4"> From Table 2 and Table 3, we can see that the system accuracy is improved by the global word sense disambiguation. The improvement is not big, but there are no analysis result that changed for the worse. The improvement is hardly contributed by the verb sense disambiguation, but mainly by the noun sense disambiguation. This is because appropriate case frames are used in many cases without the verb sense disambiguation, and this process did not lead to the improvement. The number of case frames of the verbs to which the verb sense disambiguation is applied decreased to 16%, and this indicates that the analysis efficiency improved significantly. In addition, we evaluated randomly selected 100 nouns to which the noun sense disambiguation is applied. 91 nouns were disambiguated correctly, and quite high disambiguation accuracy was achieved.</Paragraph>
  </Section>
class="xml-element"></Paper>