pdf

Chinese word segmentation and part-of-speech tagging integrated system based on undirected graph sequence labeling model

  • 2013-09-22
  • 252.51KB
  • Points it Requires : 2

           In Chinese lexical analysis, word segmentation is a necessary stage for part-of-speech tagging. In order to fully utilize the information of part-of-speech tagging and reduce the accumulation of errors in the two stages at the word segmentation stage, the best way is to integrate the two stages into one architecture. Based on the undirected graph model, this paper organically unifies word segmentation and part-of-speech tagging in a sequence tagging model. Since deeper dependencies can be used as features, the integrated system achieved a word segmentation accuracy of 97.19% and a part-of-speech tagging accuracy of 95.34% on the 1998 People\'s Daily corpus, which is the best result achieved by similar systems on this corpus.

unfold

You Might Like

Uploader
hellopinkgirls
 

Recommended ContentMore

Popular Components

Just Take a LookMore

EEWorld
subscription
account

EEWorld
service
account

Automotive
development
circle

About Us Customer Service Contact Information Datasheet Sitemap LatestNews


Room 1530, 15th Floor, Building B, No.18 Zhongguancun Street, Haidian District, Beijing, Postal Code: 100190 China Telephone: 008610 8235 0740

Copyright © 2005-2024 EEWORLD.com.cn, Inc. All rights reserved 京ICP证060456号 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号
×