<?xml version="1.0" standalone="yes"?>
<Paper uid="P98-1020">
  <Title>Trigger-Pair Predictors in Parsing and Tagging</Title>
  <Section position="10" start_page="136" end_page="136" type="concl">
    <SectionTitle>
7 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper, we have shown that, as in the case of words, there is a substantial amount of information outside the sentence which could be used to supplement tagging and parsing models. We have also shown that knowledge of the type of document being processed greatly increases the usefulness of triggers.</Paragraph>
    <Paragraph position="1"> If this information is known, or can be predicted accurately from the history of a given document being processed, then model interpolation techniques (Jelinek et al., 1980) could be employed, we anticipate, to exploit this to useful effect.</Paragraph>
    <Paragraph position="2"> Future research will concentrate on incorporating trigger-pair information, and extrasentential information more generally, into more sophisticated models of parsing and tagging. An obvious first extention to this work, for the case of tags, will be, following (Rosenfeld, 1996), to incorporate the triggers into a maximum-entropy model using trigger pairs in addition to unigram, bigram and trigram constraints.</Paragraph>
    <Paragraph position="3"> Later we intend to incorporate trigger information into a probabilistic English parser/tagger which is able to ask complex, detailed questions about the contents of a sentence. From the results presented here we are optimistic that the additional, extrasentential information provided by trigger pairs will benefit such parsing and tagging systems.</Paragraph>
  </Section>
class="xml-element"></Paper>