<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-1030">
  <Title>Automated Japanese Essay Scoring System based on Articles Written by Experts</Title>
  <Section position="9" start_page="238" end_page="239" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> An automated Japanese essay scoring system called Jess has been created for scoring essays in college-entrance exams. This system has been shown to be valid for essays of 800 to 1,600 characters. Jess, however, uses editorials and columns taken from the Mainichi Daily News newspaper as learning models, and such models are not sufficient for learning terms used in scientific and technical fields such as computers. Consequently, we found that Jess could return a low evaluation of &amp;quot;content&amp;quot; even for an essay that responded well to the essay prompt. When analyzing content, a mechanism is needed for automatically selecting  a term-document cooccurrence matrix in accordance with the essay targeted for evaluation. This enable the users to avoid reverse-engineering that poor quality essays would produce perfect scores, because thresholds for detecting the outliers on rhetoric features may be varied.</Paragraph>
  </Section>
class="xml-element"></Paper>