<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-1216">
  <Title>Named Entity Recognition in Biomedical Texts using an HMM Model</Title>
  <Section position="6" start_page="121" end_page="121" type="evalu">
    <SectionTitle>
5 Experiment Result
</SectionTitle>
    <Paragraph position="0"> The experiment results are shown in Table 3:  The Baseline2 outperforms Baseline1 because it prevents from using low frequency unigrams, and our system outperforms Baseline1 and Baseline2 because it prevents from using low frequency bigrams and unigrams. Our system benefits from huge unlabeled corpus.</Paragraph>
  </Section>
class="xml-element"></Paper>