An improved model of MST for Chinese dependency parsing

Author

Guiping Zhang ; Yan Wang ; Duo Ji

Author_Institution

Res. Center for Knowledge Eng., Shen Yang Aerosp. Univ., Shen Yang, China

fYear

2012

fDate

Oct. 30 2012-Nov. 1 2012

Firstpage

1454

Lastpage

1458

Abstract

In this paper, a Chinese dependency parsing method is proposed based on improved Maximum Spanning Tree (MST) Parser. Within this method, dependency direction discrimination model and head POS recognition model are used to modify the weights of directed edges in the MST model, and then the Eisner algorithm is used to search and generate the dependency trees. In this paper, the problems of dependency direction discrimination and head POS recognition are converted into sequence labeling; and the modeling is done by condition random fields. We tested our method on CoNLL 2009 Share Task, and the Unlabeled Attachment Score reached 86.27%.

Keywords

directed graphs; grammars; natural language processing; random processes; text analysis; trees (mathematics); Chinese dependency parsing method; CoNLL 2009 Share Task; Eisner algorithm; condition random fields; dependency direction discrimination model; directed edge weights; head POS recognition model; improved MST model; improved maximum spanning tree parser; sequence labeling; unlabeled attachment score; Accuracy; Algorithm design and analysis; Analytical models; Grammar; Labeling; Magnetic heads; Training; condition random fields; dependency parsing; maximum spanning tree;

fLanguage

English

Publisher

ieee

Conference_Titel

Cloud Computing and Intelligent Systems (CCIS), 2012 IEEE 2nd International Conference on

Conference_Location

Hangzhou

Print_ISBN

978-1-4673-1855-6

Type

conf

DOI

10.1109/CCIS.2012.6664626

Filename

6664626