<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-2077">
  <Title>Reinforcing English Countability Prediction with One Countability per Discourse Property</Title>
  <Section position="7" start_page="601" end_page="601" type="concl">
    <SectionTitle>
6 Conclusions
</SectionTitle>
    <Paragraph position="0"> This paper has proposed a method for reinforcing English countability prediction by introducing one countability per discourse. The experiments have shown that the proposed method successfully overrode original mispredictions using ef ciently the one countability per discourse property. They also have shown that it outperformed other methods used for comparison. From these results, we conclude that the proposed method is effective in reinforcing English countability prediction.</Paragraph>
    <Paragraph position="1"> In addition, the proposed method has two advantages. The rst is its applicability. It can reinforce almost any earlier method. Even to hand-coded rules, it can be applied as long as they give predictions with their con dences. This further gives an additional advantage. Recall that the instances tagged with ? by the tagging rules are discarded when training data are generated as described in Subsection 4.1. These instances can be retagged with their countability by using the proposed method and some kind of bootstrapping (Yarowsky, 1995). This means increase in training data, which might eventually result in further improvement. The second is that the proposed method is unsupervised. It requires no human intervention to reinforce countability prediction.</Paragraph>
    <Paragraph position="2"> For future work, we will investigate what models are most appropriate for exploiting the one countability per discourse property. We will also explore a method for including instances tagged with ? in training data by using the proposed method and bootstrapping.</Paragraph>
  </Section>
class="xml-element"></Paper>