论文信息 - Sequences of Part of Speech Tags vs. Sequences of Phrase Labels: How Do They Help in Parsing?

Sequences of Part of Speech Tags vs. Sequences of Phrase Labels: How Do They Help in Parsing?

We compare the contributions made by sequences of part of speech tags and sequences of phrase labels for the task of grammatical relation finding. Both are used for grammar induction, and we show that English labels of grammatical relations follow a very strict sequential order, but not as strict as POS tags, resulting in better performance of the latter on the relation finding task.

Maarten de Rijke | Gabriel G. Infante López

[1] Hinrich Schütze,et al. Book Reviews: Foundations of Statistical Natural Language Processing , 1999, CL.

[2] Maarten de Rijke,et al. Alternative approaches for Generating Bodies of Grammar Rules , 2004, ACL.

[3] Michael Collins,et al. Three Generative, Lexicalised Models for Statistical Parsing , 1997, ACL.

[4] S. Buchholz,et al. Memory-Based Grammatical Relation Finding , 2002 .

[5] Michael Collins,et al. Head-Driven Statistical Models for Natural Language Parsing , 2003, CL.

[6] Fernando Pereira,et al. Relating Probabilistic Grammars and Automata , 1999, ACL.

[7] G. G. Infante Lopez,et al. Two-level probabilistic grammars for natural language parsing , 2005 .

[8] Jason Eisner,et al. Three New Probabilistic Models for Dependency Parsing: An Exploration , 1996, COLING.

[9] Eugene Charniak,et al. Tree-Bank Grammars , 1996, AAAI/IAAI, Vol. 2.

[10] Michael Collins,et al. Three Generative, Lexicalised Models for Statistical Parsing , 1997, ACL.

[11] Colin de la Higuera,et al. Probabilistic DFA Inference using Kullback-Leibler Divergence and Minimality , 2000, ICML.