Xiangyu Duan, Jun Zhao, Bo Xu
Deterministic dependency parsing has often been regarded as an efficient parsing algorithm while its parsing accuracy is a little lower than the best results reported by more complex parsing models. In this paper, we compare deterministic dependency parsers with complex parsing methods such as generative and discriminative parsers on the standard data set of Penn Chinese Treebank. The results show that, for Chinese dependency parsing, deterministic parsers outperform generative and discriminative parsers. Furthermore, basing on the observation that deterministic parsing algorithms are greedy algorithms which choose the most probable parsing action at every step, we propose three kinds of ungreedy deterministic dependency parsing algorithms to globally model parsing actions. We take the original deterministic dependency parsers as baseline systems. Results show that ungreedy deterministic dependency parsers perform better than the baseline systems while maintaining the same time complexity, and the best result improves much over baseline.
Subjects: 13. Natural Language Processing; 13.3 Syntax
Submitted: Apr 5, 2007