<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0706">
  <Title>Using word similarity lists for resolving indirect anaphora</Title>
  <Section position="9" start_page="0" end_page="0" type="concl">
    <SectionTitle>
9 Concluding remarks
</SectionTitle>
    <Paragraph position="0"> We tested the use of word similarity lists on resolving indirect anaphoras on Portuguese newspaper texts. We presented our heuristic for searching word similarity lists to be able to find the relation between an anaphor and its antecedent. We considered similarity lists containing proper names and lists containing just common nouns. Our heuristic was able to resolve 33.3% of the cases, with precision of 52.7% when considering just common nouns, and we got 22.1%recall with precision of 36.8% when including proper names. Even though considering proper names give us the possibility of treating more anaphora cases, we got lower precision than using the lists with only nouns, since such lists are more homogeneous. These results are comparable to previous work dealing with such complex anaphora.</Paragraph>
    <Paragraph position="1"> As future work, we intend to integrate our heuristic for indirect anaphora with other heuristics for anaphora resolution into ART and investigate the best combination of application of these. Concerning refining the proposed heuristic, we intend to run more experiments aiming to tune the penalising weights when choosing an antecedent among the candidates already selected by the search on the similarity lists.</Paragraph>
  </Section>
class="xml-element"></Paper>