摘要
[目的/意义]词汇链是文本中一系列词汇关联而成的语义链。构造词汇链有助于读者把握文本主题,对知识元构建、自动文摘生成等领域研究有重要价值。[方法/过程]借鉴HNC理论的概念层次原理对词汇语义进行形式化描述,利用HNC的符号和同行优先准则,及依存句法揭示的多义词上下文信息,对词语语义进行消歧处理,进而运用HNC语义相似度计算方法获得词语的语义关联性。[结果/结论]在此基础上,提出词汇链构建算法,并对词汇链进行优选处理,获得优选词汇链。实验结果表明,基于HNC理论和依存句法提出的词汇链构造方法的可接受度较高。
[ Purpose/Significance] The lexical chain is a semantic chain connected with a series of words in the text. Building lexical chains helps readers to grasp the theme of the text, and has an important value on knowledge element building and automatic abstract generation. [ Method/Process] Based on the HNC （ Hierarchical Network of Concepts） theory, this paper describes the lexical semantics for- mally, using the symbols and the same line priority criteria of the HNC, and the polysemy contextual information revealed by dependency syntax. [ Result/Conclusion] Then the semantic relatedness of words is obtained by applying HNC semantic similarity calculation. On this basis, a lexical chain building algorithm is proposed, and the preferred lexical chain is achieved after optimization processing. Experiment results show that the proposed lexical chain building method is of a high acceptance degree.
出处
《情报杂志》
CSSCI
北大核心
2016年第2期182-187,共6页
Journal of Intelligence
关键词
词汇链
语义计算
HNC理论
依存句法
lexical chain semantic calculation HNC theory dependency syntax
