<?xml version="1.0" standalone="yes"?>
<Paper uid="W97-0208">
  <Title>Sense Tagging: Semantic Tagging with a Lexicon</Title>
  <Section position="9" start_page="50" end_page="50" type="evalu">
    <SectionTitle>
8 Results
</SectionTitle>
    <Paragraph position="0"> We have conducted some preliminary testing of this approach: our tests were run on 10 handdisambiguated sentences from the Wall Street Journal amounting to a 209 word corpus. We found that of, the word tokens which had more than 1 homograph, 86% were assigned the correct homo.</Paragraph>
    <Paragraph position="1"> graph and 57% of tokens were assigned the correct sense using our simple tagger. These figures should be compared to 72% correct homograph assignment and 47% correct sense assignment using simulated annealing alone on the same test set (see (Cowie, Guthrie, and Guthrie, 1992)). It should he noted that the granularity of sense distinctions at the LDOCE homograph level (eg. &amp;quot;bank&amp;quot; as 'edge of river' or 'financial institution') is the same as the distinctions made by current small-scale WSD algoritbm~ (eg. (Gale, Church, and Yarowsky, 1992), (Yarowsky, 1993), (Schfitze, 1992)) and our system is a true tagging algorithm, operating on free text.</Paragraph>
    <Paragraph position="2"> Our evaluation is unsatisfactory due to the small test set, but does demonstrate that the use of independent knowledge sources leads to an improvement in the quality of disambignation. We fully expect our results to improve with the addition of further, independent, modules.</Paragraph>
  </Section>
class="xml-element"></Paper>