<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-0118">
  <Title>Voting between Dictionary-based and Subword Tagging Models for Chinese Word Segmentation</Title>
  <Section position="7" start_page="128" end_page="128" type="concl">
    <SectionTitle>
5 Conclusion
</SectionTitle>
    <Paragraph position="0"> Our Chinese word segmentation system is based on majority voting among the initial outputs from forward maximum matching, from a CRF model with maximum subword-based tagging, and from a CRF model with minimum subword-based tagging. In addition, we experimented with various steps in post-processing which effectively boosted the overall performance.</Paragraph>
    <Paragraph position="1"> In future research, we shall explore more sophisticated ways of voting, including the continuing investigation on the segmentation lattice approach. Also, more powerful methods on how to accurately deal with unknown words, including person and place names, without external knowledge, will be studied as well.</Paragraph>
  </Section>
class="xml-element"></Paper>