<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-3408">
  <Title>ChAT: A Time-Linked System for Conversational Analysis</Title>
  <Section position="5" start_page="55" end_page="55" type="concl">
    <SectionTitle>
4 Discussion
</SectionTitle>
    <Paragraph position="0"> In this paper, we have presented ChAT, a system designed to aid in the analysis of any kind of threaded dialogue. Our system is designed to be flexible in that the UI and processing components work with multiple data types. The processing components can be used independently, or within the UI. The UI aids users in in-depth analysis of individual conversations. The components can be run independent of the UI and in batch, resulting in an xml document containing the original transcripts and the metadata added by the processing components. This functionality allows the data to be manipulated by traditional text mining techniques, or to be viewed in any other visualization. We have not performed user evaluation on the interface. Our topic segmentation performs better than the current state of the art, and named-entity extraction we have integrated is commercial grade.</Paragraph>
    <Paragraph position="1"> We are currently working on an evaluation of the affect scoring. While our topic segmentation is good, we are working to improve the labels we use for the topics. Most importantly, we plan on addressing the usefulness of the UI with user studies.</Paragraph>
  </Section>
class="xml-element"></Paper>