计算机工程与应用 ›› 2014, Vol. 50 ›› Issue (8): 93-98.

• 数据库、数据挖掘、机器学习 • 上一篇    下一篇

现代汉语“对象格”自动识别研究

汪梦翔1,2,王厚峰1,2,刘  扬1,2,张龙凯1   

  1. 1.北京大学 计算语言学研究所,北京 100871
    2.北京大学 计算语言学教育部重点实验室,北京 100871
  • 出版日期:2014-04-15 发布日期:2014-05-30

Study for automatic recognition of Chinese object target role

WANG Mengxiang1,2, WANG Houfeng1,2, LIU Yang1,2, ZHANG Longkai1   

  1. 1.Institute of Computational Linguistics, Peking University, Beijing 100871, China
    2.Key Lab of Computational Linguistics, MoE, Peking University, Beijing 100871, China
  • Online:2014-04-15 Published:2014-05-30

摘要: 在中文语义角色标注体系中,“对象格”作为汉语语法体系的“自产物”,有着区别于其他语义角色的句法和语义的特征。讨论“对象格”的语义特征及相关语义角色标注间的纠结现象,提出了面对自然语言处理的“对象”语义角色标注流程图,并进行了相关实验验证。

关键词: 对象, 语义角色, 自动标注

Abstract: In the Chinese semantic role labeling system, as a “native product” in the Chinese grammar, “Object Target” role has syntactic and semantic features. This paper discusses the semantic features of “Object Target” and the tangled problems in related semantic role labeling. It puts forward a flow chart for the machine recognition of Chinese “Object Target” role and makes an experimental demonstration.

Key words: object target, semantic role, automatic tagging