<?xml version="1.0" standalone="yes"?>
<Paper uid="W98-1123">
  <Title>Linear Segmentation and Segment Significance</Title>
  <Section position="3" start_page="203" end_page="203" type="relat">
    <SectionTitle>
4 Future Work
</SectionTitle>
    <Paragraph position="0"> Improvements to the current system can be categorized along the lines of the two modules.</Paragraph>
    <Paragraph position="1"> For segmentation, applying machine learning techniques (Beeferman et al. 1997) to learn weights is a high priority. Moreover we feel shared resources for segmentation evaluation should be established', to aid in a comprehensive cross-method study and to help alleviate the problems of significance of small-scale evaluations as discussed in Klavans et al (1998). ' For the purposes of our own evaluation, we constructed web-based software tool that allows users to annotate a document with segmentation markings. We propose initiating a distributed cross evaluation of text segmentation work, using our system as a component to store and share user-given and automatic markings.</Paragraph>
    <Paragraph position="2"> For judging segment function, we plan to perform a direct assessment of the accuracy of segment classification. We want to expand and ref'me our definition of the types of segment function to include more distinctions, such as the difference between document/segment borders (Reynar 1994). This would help in situations where input consists of multiple articles or a continuous stream, as in Kanade et al. (1997).</Paragraph>
  </Section>
class="xml-element"></Paper>