基于依存树库的弹幕语言句法特征研究
Research on the Syntactic Features of Bullet Screen Language Based on Dependency Tree Library
王紫悦 1阿孜古丽·夏力甫1
作者信息
- 1. 新疆大学 中国语言文学学院,新疆 乌鲁木齐 830002
- 折叠
摘要
基于哈尔滨工业大学LTP语言云系统的依存句法树库,从依存关系类型、依存距离、依存方向百分比三个方面,对弹幕语言的句法特征进行探究.研究发现,弹幕语言的依存类型以状中关系、主谓关系、右附加关系、定中关系、动宾关系为主.弹幕语言的句长较短,以简单句为主,复句为辅;单句的句长集中在1-5 个词语之内,复句的句长集中在 10-15 个词语之内.弹幕语言单、复句的平均依存距离随着句长的增加而缓慢增长,二者具有显著相关性.弹幕语言的依存方向主要表现为支配词在后.
Abstract
Based on the dependency syntax tree library of LTP language Cloud system of Harbin Institute of Technology,this paper explores the syntactic features of bullet screen language from three aspects:dependency type,dependency distance and dependency direction percentage.It is found that the dependency types of bullet screen language are mainly in the form of the relationship,the subject-predicate relationship,the right attachment relationship,the fixed middle relationship and the verb-object relationship.The sentences of bullet screen language are short,mainly simple sentences,supplemented by complex sentences;The sentence length of a single sentence is concentrated within 1-5 words,and the sentence length of a complex sentence is concentrated within 10-15 words.The average dependency distance of single complex sentences in bullet screen language increases slowly with the increase of sentence length,and there is a significant correlation between them.The dependent direction of bullet screen language is mainly manifested as the dominant word after.
关键词
弹幕语言/依存树库/依存关系类型/依存距离/依存方向/句法特征Key words
bullet screen language/dependency tree library/dependency type/dependency distance/dependency direction/syntactic feature引用本文复制引用
出版年
2024
NETL
NSTL
万方数据