摘要
语义相关度计算是智能信息处理和文本挖掘领域的重要研究内容。本文将维基百科作为语义知识库,利用维基百科层次分类体系结构、解释页面之间的链接结构等结构特征进行词语之间的相关度计算。针对层次分类体系的有向无环图结构,采用多路径语义相关度计算方法进行相关度计算;针对链接的重要性和类型,融合链接权重和链接类型进行相关度计算。实验结果表明,该方法取得了预期的实验效果,证明该方法是可行和有效的。
Semantic relatedness computing is one of the key research content in intelligent informationprocessing and text mining. Based on Wikipedia, this paper computes the relatedness between words withtaxonomic hierarchies and link structure. Adopts multipath semantic relatedness calculation method forthe acyclic directed graph structure of taxonomic hierarchies, combines link weight and link type for theimportance and types of link structure. Experiment results demonstrate that this method achieved a goodanticipative effect, and this method is feasible and effective.
出处
《情报科学》
CSSCI
北大核心
2015年第9期72-75,120,共5页
Information Science
关键词
语义相关度
结构特征
维基百科
semantic relatedness
structure feature
wikipedia