<?xml version="1.0" standalone="yes"?>
<Paper uid="W00-0720">
  <Title>Genetic Algorithms for Feature Relevance Assignment in Memory-Based Language Processing</Title>
  <Section position="5" start_page="104" end_page="105" type="concl">
    <SectionTitle>
3 Conclusions and Related Research
</SectionTitle>
    <Paragraph position="0"> The issue of feature-relevance assignment is well-documented in the machine learning literature. Excellent comparative surveys are (Wettschereck, Aha, and Mohri, 1997) and (Wettschereck and Aha, 1995) or (Blum and Langley, 1997). Feature subset selection by means of evolutionary algorithms was investigated by Skalak (1994), Vafaie and de Jong (1992), and Yang and Honavar (1997).</Paragraph>
    <Paragraph position="1"> Other work deals with evolutionary approaches for continuous feature weight assignment such as Wilson and Martinez (1996), or Punch and Goodman (1993).</Paragraph>
    <Paragraph position="2"> The conclusions from these papers are in agreement with our findings on the natural language data, suggesting that feature selection and weighting with GA's significantly outperform non-weighted approaches. Feature selection generally improves accuracy with a reduc6This fits in with current theory about this morphological process (e.g. Trommelen (1983), Daelemans et al. (1997)).</Paragraph>
    <Paragraph position="3">  tion in the number of features used. However, we have found no results (on these particular data) that indicate an advantage of evolutionary feature selection approach over the more classical iterative methods. Our experiments further show that there is no evidence that GA weighting is in general competitive with simple filter methods such as gain ratio. Possibly, a parameter setting for the GA could be found that gives better results, but searching for such an optimal parameter setting is at present computationally unfeasible for typical natural language processing problems.</Paragraph>
  </Section>
class="xml-element"></Paper>