<?xml version="1.0" standalone="yes"?>
<Paper uid="I05-3028">
  <Title>Chinese Word Segmentation with Multiple Postprocessors in HIT-IRLab</Title>
  <Section position="8" start_page="174" end_page="174" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> Through the second SIGHAN bakeoff, we find the segmentation model and the algorithm in our system is effective and the multiple postprocessors we use can also enhance the performance of our system. At the same time, we also find some problems of us. It also has potential for us to improve our system. Take MDW for example, we can make use of more features such as the POS and the length of the preceding word to enhance the recall and precision rate. The bakeoff points out the direction for us to improve our system.</Paragraph>
  </Section>
class="xml-element"></Paper>