<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-1305">
  <Title>Two-Phase Biomedical NE Recognition based on SVMs</Title>
  <Section position="8" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper, we proposed a new method of two-phase biomedical named entity recognition based on SVMs and dictionary-lookup. At the first phase, we tried to identify each entity with one SVM classifier and to post-process with a simple dictionary look-up for correcting the errors by the SVM. At the second  phase, we tried to classify the identified entity into its semantic class by voting the SVMs. By dividing the task into two subtasks, the identification and the semantic classification task, we could select more relevant features for each task and take an alternative classification method according to the task. This is resulted into the mitigation effect of the unbalanced class distribution problem but also improvement of the performance of the overall tasks.</Paragraph>
  </Section>
class="xml-element"></Paper>