<?xml version="1.0" standalone="yes"?>
<Paper uid="W97-0318">
  <Title>Learning Methods for Combining Linguistic Indicators to Classify Verbs</Title>
  <Section position="6" start_page="159" end_page="160" type="concl">
    <SectionTitle>
5 Conclusions and Future Work
</SectionTitle>
    <Paragraph position="0"> We have compiled a set of fourteen quantitative linguistic indicators that, when used together, significantly improve the classification of verbs according to stativity. The values of these indicators are measured automatically across a corpus of text.</Paragraph>
    <Paragraph position="1"> Each of three machine learning techniques successfully combined the indicators to improve classification performance. The best of the three, decision tree induction, achieved a classification accuracy of 93.9%, as compared to the uninformed baseline's accuracy of 83.8%. Furthermore, genetic programming and log-linear regression also achieved improvements over the baseline. These results were measured over an unrestricted set of verbs.</Paragraph>
    <Paragraph position="2"> The improvement in classification performance is more dramatically illustrated by the favorable trade-off between stative and event recall achieved by all three of these methods, which is profitable for tasks that weigh the identification of states more heavily than events.</Paragraph>
    <Paragraph position="3"> This analysis has revealed correlations between stativity and five indicators that are not traditionally linked to stativity in the linguistic literature. Furthermore, one of these four, verb frequency, individually increased classification accuracy from the baseline method to 88.0%.</Paragraph>
    <Paragraph position="4"> To classify a clause, the current system uses only the indicator values corresponding to the clause's main verb. This procedure could be expanded to  incorporate rules that classify a clause directly from clausal features (e.g., Is the main verb show, is the clause in the progressive?), or by calculating indicator values over other clausal constituents in addition to the verb (Siegel and McKeown, 1996; Siegel, 1997).</Paragraph>
    <Paragraph position="5"> Classification performance may also improve by incorporating additional linguistic indicators, such as co-occurrence with rate adverbs, e.g., quickly, or occurrences as a complement of force or persuade, as suggested by Klavans and Chodorow (1992).</Paragraph>
  </Section>
class="xml-element"></Paper>