<?xml version="1.0" standalone="yes"?>
<Paper uid="J02-4006">
  <Title>Using Hidden Markov Modeling to Decompose Human-Written Summaries</Title>
  <Section position="10" start_page="542" end_page="542" type="concl">
    <SectionTitle>
7. Conclusions
</SectionTitle>
    <Paragraph position="0"> We defined the problem of decomposing human-written summaries and proposed a hidden Markov model solution to the problem. The decomposition program can automatically determine whether a summary sentence is constructed by reusing text from the original document; it can accurately recognize the reused phrases in a summary sentence despite their different granularities; it can also pinpoint the exact origin in the document for a phrase. The algorithm is fast and straightforward. It does not need other tools such as a tagger or parser as preprocessors. It does not have complex processing steps. The evaluations show that the program performs very well for the decomposition task.</Paragraph>
  </Section>
class="xml-element"></Paper>