<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-1609">
  <Title>An Unsupervised Approach for Bootstrapping Arabic Sense Tagging</Title>
  <Section position="9" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusions
</SectionTitle>
    <Paragraph position="0"> We presented, SALAAM, a method for bootstrapping the sense disambiguation process for Arabic texts using an existing English sense inventory leveraging translational correspondence between Arabic and English. SALAAM achieves an absolute precision of 56.9% on the task for Arabic. Of the 673 correctly tagged English tokens for the SENSEVAL2 English All Words Task, approximately 90% of the Arabic data is deemed correctly tagged by 3 native speakers of Arabic. Therefore, SALAAM is validated as a very good first approach to Arabic WSD. Moreover, we perform a preliminary investigation with very promising results into the quality of the English sense inventory, WN1.7pre, as an approximation to an Arabic sense inventory.</Paragraph>
  </Section>
class="xml-element"></Paper>