摘要
针对概念语义相似度计算的片面性和不完善性问题,该文提出了一种综合的概念语义相似度计算方法。首先,基于传统的语义距离模型,综合考虑了有向边的类型、深度、密度、方向等因素,计算概念相似度;然后,根据本体的五元组组成元素,分别计算概念相似度;最后,使用sigmoid阈值函数求得各个相似度对应的调节因子值,并进行综合,实现了调节因子与各个相似度的自适应。实验证明,该方法有效且精确。
To aim at the one-sided and incomplete issues of the computation of concept similarity,this paper puts forward a compositive approach.Firstly,based on the traditional semantic distance,considering the type of directed edges-"IS-A"category relations,density,direction and other factors,it can compute a similarity.Then,according to the elements of the ontology five-tuple,it can calculate the similarity of concepts.Finally,in order to realize the adaptability of each similarity with the regulatory factor,it use the sigmoid threshold function to calculate the value of the regulatory factor corresponding to each similarity,and then calculate comprehensively.Experiments show that the method is effective and accurate.
出处
《电脑知识与技术》
2011年第1X期619-622,共4页
Computer Knowledge and Technology
关键词
本体
语义相似度
语义距离
本体五元组
调节因子
ontology
semantic similarity
semantic distance
ontology five-tuple
regulatory factor
