<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-1026">
  <Title>HowtogetaChineseName(Entity): Segmentation and Combination Issues</Title>
  <Section position="9" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper, we discuss two topics related to Chinese NE recognition: dealing with language-specific issues such as word segmentation, and combining multiple classifiers to enhance the system performance. In the described experiments, the character-based model consistently outperforms the word-based model - one major reason for this fact is that the segmentation granularity might not be suited for this particular task. Combining four statistical classifiers, including a hidden Markov model classifier, a transformation-based learning classifier, a maximum entropy classifier, and a robust risk minimization classifier, significantly improves the system performance, yielding a 10% relative reduction in F-measure error over the best performing classifier.</Paragraph>
  </Section>
class="xml-element"></Paper>