<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-1112">
  <Title>Chinese Term Extraction from Web Pages Based on Compound word Productivity</Title>
  <Section position="7" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper, we apply automatic term recognition system based on FLR proposed by Nakagawa and Mori (2003) to Chinese Web pages because the term extraction from small text like one Web page is the future oriented topic. We proposed two methods: word based and character based extraction and ranking using the compound word productivity of simple words. Since the accuracies of term recognition are around 60% for top 1,000 term candidates in NTCIR TMREC task(Kageura et al 1999), the result of 75% accuracy of top ten candidates is a good start.</Paragraph>
  </Section>
class="xml-element"></Paper>