论文信息 - A Novel Probabilistic Model for Dependency Parsing

A Novel Probabilistic Model for Dependency Parsing

A new knowledge based probabilistic dependency parsing (KPDP) is presented to compensate the local optimization problem of native probabilistic models. KPDP is composed of two stages: (1) selecting a set of constituent parse trees with an extensive bottom-up chart parsing algorithm which employs Maximum Entropy Models to calculate single arc probabilities; (2) finding the best parsing tree with the help of word knowledge. Based on case grammar theory, the word knowledge is represented as some patterns which group those arcs with the same head. KPDP is evaluated experimentally using the dataset distributed in CoNLL 2008 share-task. An unlabelled arc score of 87 % is reported, which is 3.39% higher than the native model without word knowledge. This work will contribute to and stimulate other researches in the field of parsing.

Xiaolong Wang | Xuan Wang | Shixi Fan

[1] Fernando Pereira,et al. Online Learning of Approximate Dependency Parsing Algorithms , 2006, EACL.

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

[3] Adam L. Berger,et al. A Maximum Entropy Approach to Natural Language Processing , 1996, CL.

[4] Qin Iris Wang,et al. Learning Structured Classifiers for Statistical Dependency Parsing , 2007, NAACL.

[5] Yuji Matsumoto,et al. Statistical Dependency Analysis with Support Vector Machines , 2003, IWPT.

[6] Kun Yu,et al. Chinese Dependency Parsing with Large Scale Automatically Constructed Case Structures , 2008, COLING.

[7] Eugene Charniak,et al. A Maximum-Entropy-Inspired Parser , 2000, ANLP.

[8] Lucien Tesnière. Éléments de syntaxe structurale , 1959 .

[9] Lilja Øvrelid,et al. Linguistic features in data-driven dependency parsing , 2008, CoNLL.

[10] Daisuke Kawahara,et al. Minimally Lexicalized Dependency Parsing , 2007, ACL.

[11] Fernando Pereira,et al. Non-Projective Dependency Parsing using Spanning Tree Algorithms , 2005, HLT.

[12] Joakim Nivre,et al. Characterizing the Errors of Data-Driven Dependency Parsing Models , 2007, EMNLP.