摘要
指代的解决对于自然语言理解的非歧义具有重要的意义重视篇章和故事情节的理解固然有助于解决类似的二义性问题，但所需的巨大知识库在实际系统中，无论是建立还是最终的效率和完备性均很难满足用户的需要．文章以第三人称代词为代表，分析其各种指代现象并进行分类通过语法分析和句子语义平面的主客观信息研究，利用格框架，提出了在上下文相关的语义环境中进行指代分类解决的思想，并给出了相应算法．
Anaphora resolution is very important for the unambiguity understaning of natural language. Certainly, good Understanding of text and plot is helPful for it, but it needs a huge knowledge base. In practice, it is very difficult to create a complete base that could satisfy the users' needs. The anaphora resolution (AR) of third person Pronounce is mainly discussed. Based on closer examination. The anaphora are classified into severa kinds. With the research on the subjective and obective information of Chinese sentences on the semantic level and the study of case frame, a new way of anaphora resolution is provided. corresponding algorithms are also given.
出处
《西南师范大学学报（自然科学版）》
CAS
CSCD
北大核心
1999年第6期633-637,共5页
Journal of Southwest China Normal University(Natural Science Edition)
