<?xml version="1.0" standalone="yes"?>
<Paper uid="H05-1065">
  <Title>Proceedings of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing (HLT/EMNLP), pages 515-522, Vancouver, October 2005. c(c)2005 Association for Computational Linguistics Disambiguation of Morphological Structure using a PCFG</Title>
  <Section position="10" start_page="521" end_page="521" type="concl">
    <SectionTitle>
9 Summary
</SectionTitle>
    <Paragraph position="0"> We presented a disambiguation method for German morphological analyses which is based on a head-lexicalized probabilistic context-free grammar. The words are split into morpheme lattices by a finite state morphology, and then parsed with the probabilistic context-free grammar. The grammar was trained on unlabeled data using the Inside-Outside algorithm and evaluated on 807 manually disambiguated analyses of infrequent words. Different training strategies have been compared. A combination of one iteration of unlexicalized training and four iterations of lexicalized training returned the best results with over 68% exact match accuracy, compared to a baseline of 45% which was obtained by randomly choosing one of the minimally complex analyses. Without lexicalization, the accuracy dropped by 15 percentage points, indicating that lexicalization is essential for morphological disambiguation. null</Paragraph>
  </Section>
class="xml-element"></Paper>