<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-1101">
  <Title>Segmentation of Chinese Long Sentences Using Commas</Title>
  <Section position="8" start_page="7" end_page="7" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper, we propose a method to segment a Chinese sentence by classification of the comma.</Paragraph>
    <Paragraph position="1"> We define the criteria for classification, and according to the criteria, a model for classification of the comma is given. The segmentation at the comma position seems to be efficient for improving the accuracy of dependency parsing by 9.6percent. Moreover, since commas more frequently appear in Chinese language, we expect our approach including salient and refined analysis of comma usages provides feasible solutions for segmentation. null However, the accuracy for the segmentation is not yet satisfactory. Since erroneous segmentation may cause a parsing failure for the entire sentence, errors can be serious. Further research should be done to improve the performance and reduce side effects for parsing the entire sentence.</Paragraph>
  </Section>
class="xml-element"></Paper>