<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-1643">
  <Title>Sydney, July 2006. c(c)2006 Association for Computational Linguistics A Skip-Chain Conditional Random Field for Ranking Meeting Utterances by Importance[?]</Title>
  <Section position="11" start_page="371" end_page="371" type="concl">
    <SectionTitle>
9 Conclusion
</SectionTitle>
    <Paragraph position="0"> An order-2 CRF with skip-chain dependencies derived from the automatic analysis of participant interaction was shown to outperform linear-chain BNs and CRFs, despite the incorporation in all cases of the same competitive set of predictors resulting from cross-validated feature selection.</Paragraph>
    <Paragraph position="1"> Compared to an order-0 CRF model, the absolute increase in performance is 3.9% (7.5% relative increase), which indicates that it is helpful to use skip-chain sequence models in the summarization task. Our best performing system reaches 91.3% of human performance, and scales relatively well on automatic speech recognition output.</Paragraph>
  </Section>
class="xml-element"></Paper>