<?xml version="1.0" standalone="yes"?>
<Paper uid="P01-1045">
  <Title>From Chunks to Function-Argument Structure: A Similarity-Based Approach</Title>
  <Section position="6" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusion and Future Research
</SectionTitle>
    <Paragraph position="0"> The results of 89.73 % (German) and 90.40 % (English) correctly assigned functional labels validate the general approach. We anticipate further improvements by experimenting with more sophisticated similarity metrics7 and by enriching the linguistic information in the instance base. The latter can, for example, be achieved by preserving more structural information contained in the chunk parse. Yet another dimension for experimentation concerns the way in which the algorithm generalizes over the instance base. In the current version of the algorithm, generalization heavily relies on lexical and part-of-speech information. However, a richer set of backing-off strategies that rely on larger domains of structure are easy to envisage and are likely to significantly improve recall performance.</Paragraph>
    <Paragraph position="1"> While we intend to pursue all three dimensions of refining the basic algorithm reported here, we have to leave an experimentation of which modifications yield improved results to future research. 7(Daelemans et al., 1999) reports that the gain ratio similarity metric has yielded excellent results for the NLP applications considered by these investigators.</Paragraph>
  </Section>
class="xml-element"></Paper>