For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
A New Probabilistic Dependency Parsing Model for Head-Final, Free Word Order Languages
Hoojung CHUNG Hae-Chang RIM
IEICE TRANSACTIONS on Information and Systems
Publication Date: 2003/11/01
Print ISSN: 0916-8532
Type of Manuscript: LETTER
Category: Natural Language Processing
dependency parsing, statistical parsing,
Full Text: PDF>>
We propose a dependency parsing model for head-final, variable word order languages. Based on the observation that each word has its own preference for its modifying distance and the preferred distance varies according to surrounding context of the word, we define a parsing model that can reflect the preference. The experimental result shows that the parser based on our model outperforms other parsers in terms of precision and recall rate.