<?xml version="1.0" standalone="yes"?>
<Paper uid="N06-2050">
  <Title>Comparing the roles of textual, acoustic and spoken-language features on spontaneous-conversation summarization</Title>
  <Section position="7" start_page="199" end_page="199" type="concl">
    <SectionTitle>
5 Conclusions
</SectionTitle>
    <Paragraph position="0"> Previous work on summarizing spontaneous conversations has mainly focused on textual features. This paper explores and compares both textual and speech-related features. The experiments show that these features incrementally improve summarization performance. We also find that speech disfluencies, which are removed as noise in previous work, help identify important utterances, while the structural feature is less effective than it is in broadcast news.</Paragraph>
  </Section>
class="xml-element"></Paper>