<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-1014">
  <Title>Evaluation Measures Considering Sentence Concatenation for Automatic Summarization by Sentence or Word Extraction</Title>
  <Section position="7" start_page="0" end_page="0" type="concl">
    <SectionTitle>
3 Conclusion
</SectionTitle>
    <Paragraph position="0"> We have presented the results of evaluating the appropriateness of the sentence concatenations in summaries generated using SumACCY, WSumACCY, NrstACCY and n-gram precision.</Paragraph>
    <Paragraph position="1"> We found that the measures taking into account sentence concatenation much better reflected the judgments of humans than did the single sentence precision, so the concatenation of sentences in summaries should be evaluated.</Paragraph>
    <Paragraph position="2"> Although the human judgments and the manual summaries for word extraction did not vary much among the subjects, those for sentence extraction for single article summarization greatly varied among the subjects. As a result, it is very difficult to set correct answers for single article summarization through sentence extraction.</Paragraph>
    <Paragraph position="3"> Future works involves experiments to examine the efficiency of each numerical measures in response to the coverage of correct answers.</Paragraph>
  </Section>
class="xml-element"></Paper>