<?xml version="1.0" standalone="yes"?>
<Paper uid="I05-2002">
  <Title>A Hierarchical Parsing Approach with Punctuation Processing for Long Chinese Sentences</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Until recently, although punctuations are clearly important parts of the written Chinese, many Chinese parsing systems developed to date have simply ignored them. Some researches have been done on English punctuations in parsing [1, 2, 3, 4, 5], their researches have used plenty of theoretical and experimental facts to prove that it is effective to incorporate punctuation information into parsing of long complex sentences. But as far as we know, little work has been done in Chinese syntactic parsing.</Paragraph>
    <Paragraph position="1"> Because the derivation of Chinese punctuations was referring to western language [3], they have many similarities in usage. Researches on Chinese punctuations in parsing will be valuable.</Paragraph>
    <Paragraph position="2"> However, our study shows, there are still differences between them, special research on Chinese punctuations is necessary.</Paragraph>
    <Paragraph position="3"> In this paper, differences in English and Chinese punctuations are compared and the special difficulty and corresponding cause in parsing Chinese long sentences are analyzed. Then a new hierarchical parsing (HP) approach is proposed instead of traditional parsing (TP) method. This 'divide-and-rule' strategy greatly reduces the time consumption. Open test shows, parsing accuracy and recall of HP method are both about 7% higher than those of TP.</Paragraph>
    <Paragraph position="4"> The remainder of this paper is organized as follows: Section 2 is related work. Section 3 mainly discusses the special difficulties and solution in parsing long Chinese sentences. Then HP method is discussed in detail in Section 4.</Paragraph>
    <Paragraph position="5"> Section 5 gives the final experiment results and corresponding analyses. Finally, the further work is expected.</Paragraph>
  </Section>
class="xml-element"></Paper>