<?xml version="1.0" standalone="yes"?>
<Paper uid="W99-0617">
  <Title>I POS Tags and Decision Trees for Language Modeling</Title>
  <Section position="9" start_page="135" end_page="135" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> Unlike previous approaches that use POS tags, we redefined the speech recognition problem so that it includes finding the best word sequence and best POS tag interpretation for those words. Thus this work can be seen as a first-step towards tightening the integration between speech recognition and natural language processing.</Paragraph>
    <Paragraph position="1"> In order to estimate the probabilities of our POS-based model, we use standard algorithms for clustering and growing decision trees; however, we have modified these algorithms to better use the POS information. The POS-based model results in a reduction in perplexity and in word error rate in comparison to a word-based backoff approach. Part of this improvement is due to the decision tree approach for estimating the probabilities.</Paragraph>
  </Section>
class="xml-element"></Paper>