<?xml version="1.0" standalone="yes"?>
<Paper uid="N06-2032">
  <Title>Story Segmentation of Brodcast News in English, Mandarin and Arabic</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
2 Related work
</SectionTitle>
    <Paragraph position="0"> Previous approaches to story segmentation have largely focused lexical features, such as word similarily (Kozima, 1993), cue phrases (Passonneau and Litman, 1997), cosine similarity of lexical windows (Hearst, 1997; Galley et al., 2003), and adaptive language modeling (Beeferman et al., 1999).</Paragraph>
    <Paragraph position="1"> Segmentation of stories in BN have included some acoustic features (Shriberg et al., 2000; Tcurrency1ur et al., 2001). Work on non-English BN, generally use this combination of lexical and acoustic measures, such as (Wayne, 2000; Levow, 2004) on Mandarin.</Paragraph>
    <Paragraph position="2"> And (Palmer et al., 2004) report results from feature selection experiments that include Arabic sources, though they do not report on accuracy. TRECVID has also identi ed visual cues to story segmentation of video BN (cf. (Hsu et al., 2004; Hsieh et al., 2003; Chaisorn et al., 2003; Maybury, 1998)).</Paragraph>
  </Section>
class="xml-element"></Paper>