<?xml version="1.0" standalone="yes"?>
<Paper uid="A97-1056">
  <Title>Sequential Model Selection for Word Sense Disambiguation *</Title>
  <Section position="10" start_page="393" end_page="393" type="concl">
    <SectionTitle>
8 Conclusion
</SectionTitle>
    <Paragraph position="0"> Sequential model selection is a viable means of choosing a probabilistic model to perform word-sense disambiguation. We recommend AIC as the evaluation criterion during model selection due to  the following: 1. It is difficult to set an appropriate cutoff value (a) for a significance test.</Paragraph>
    <Paragraph position="1"> 2. The information criteria AIC and BIC are more robust to changes in search strategy.</Paragraph>
    <Paragraph position="2"> 3. BIC removes too many interactions and results  in models of too low complexity.</Paragraph>
    <Paragraph position="3"> The choice of search strategy when using AIC is less critical than when using significance tests. However, we recommend FSS for sparse data (NLP data is typically sparse) since it reduces the impact of very high degrees of freedom and the resultant unreliable parameter estimates on model selection.</Paragraph>
    <Paragraph position="4"> The Naive Bayes classifier is based on a low complexity model that is shown to lead to high accuracy. If feature selection is not in doubt (i.e., it is fairly certain that all of the features are somehow relevant to classification) then this is a reasonable approach. However, if some features are of questionable value the Naive Bayes model will continue to utilize them while sequential model selection will disregard them. All of the search strategies and evaluation criteria discussed are implemented in the public domain program CoCo (Badsberg, 1995).</Paragraph>
  </Section>
class="xml-element"></Paper>