<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-1529">
  <Title>Vancouver, October 2005. c(c)2005 Association for Computational Linguistics Robust Extraction of Subcategorization Data from Spoken Language</Title>
  <Section position="5" start_page="204" end_page="204" type="concl">
    <SectionTitle>
4 Conclusion
</SectionTitle>
    <Paragraph position="0"> We showed that it should not be assumed that standard statistical parsers will fail on language that is very different from what they are trained on. Specifically, the results of Experiment 1 showed that it is feasible to apply current SCF extraction technology to spoken language. Experiment 2 showed that incorrect SCF cues due to segmentation errors and disfluency can be recognized by our linguistic heuristics. We have shown that our SCF acquisition system as a whole will work for the different demands of spoken language.</Paragraph>
  </Section>
class="xml-element"></Paper>