<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-1120">
  <Title>Accurate Collocation Extraction Using a Multilingual Parser</Title>
  <Section position="8" start_page="958" end_page="958" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> Inthispaper, weprovidedboththeoreticalandempirical arguments in the favor of performing syntactic analysis of texts prior to the extraction of collocations with statistical methods.</Paragraph>
    <Paragraph position="1"> 7A closer look at the data revealed that this might be explained by some inconsistencies between annotations.</Paragraph>
    <Paragraph position="2"> Part of the extraction work that, like ours, relies on parsing was cited in section 2. Most often, it concerns chunking rather than complete parsing; specific syntactic configurations (such as adjective-noun, preposition-noun-verb); and languages other than the ones we deal with (usually, English and German). Parsing has been also used after extraction (Smadja, 1993) for filtering out invalid results. We believe that this is not enough and that parsing is required prior to the application of statistical tests, for computing a realistic frequency profile for the pairs tested.</Paragraph>
    <Paragraph position="3"> As for evaluation, unlike most of the existing work, we are not concerned here with comparing the performance of association measures (cf.</Paragraph>
    <Paragraph position="4"> (Evert, 2004; Pecina, 2005) for comprehensive references), but with a contrastive evaluation of syntactic-based and standard extraction methods, combined with the same statistical computation.</Paragraph>
    <Paragraph position="5"> Our study finally clear the doubts on the usefulness of parsing for collocation extraction. Previous work that quantified the influence of parsing onthequalityofresultssuggestedtheperformance for tagged and parsed texts is similar (Evert and Kermes, 2003). This result applies to a quite rigid syntactic pattern, namely adjective-noun in German. But a preceding study on noun-verb pairs (Breidt, 1993) came to the conclusion that good precision can only be achieved for German with parsing. Its author had to simulate parsing because of the lack, at the time, of parsing tools for German. Our report, that concerns an actual system and a large data set, validates Breidt's finding for a new language (French).</Paragraph>
    <Paragraph position="6"> Our experimental results confirm the hypotheses put forth in section 4, and show that parsing (even if imperfect) benefits to extraction, notably by a drastic reduction of the noise in the top of the significance list. In future work, we consider investigating other levels of the significance list, extending the evaluation to other languages, comparing against shallow-parsing methods instead of the window method, and performing recall-based evaluation as well.</Paragraph>
  </Section>
class="xml-element"></Paper>