<?xml version="1.0" standalone="yes"?>
<Paper uid="W01-0909">
  <Title>A cross-comparison of two clustering methods</Title>
  <Section position="8" start_page="71" end_page="71" type="concl">
    <SectionTitle>
5 Conclusion
</SectionTitle>
    <Paragraph position="0"> We presented in this paper an approach for evaluating the results of an unsupervised learning method, when no human evaluation is possible or when no classification exists as a reference. As a result, this method builds classes of weighted words that regroup thematic units. We defined in a previous work a stability threshold of these classes, thus we aim at evaluating this subset of classes. To do that, we applied another clustering method that only needs to know the number of classes to build on the same subset of TUs and we reformulate our evaluation problem in comparing the two classifications. Our first results lead to show a great correlation between the two results.</Paragraph>
    <Paragraph position="1"> We now have to develop other tests, for example on a different number of classes, to verify our first results. A second step will be to evaluate the methods on the same task, as a classification task for example, whose protocol has to be defined.</Paragraph>
  </Section>
class="xml-element"></Paper>