基于概念间边权重的概念相似性计算方法

Concept similarity computation method based on edge weighting between concepts

介绍了传统的基于距离的相似度计算方法,针对其在距离计算中包含语义信息不充足的现状,提出了一种改进的使用WordNet的基于概念之间边的权重的相似性度量方法。该方法综合考虑了概念在词库中所处层次的深度和密度,即概念的语义丰富程度,设计了一种通用的概念语义相似性计算方法,该方法简化了传统语义相似性算法,并解决了语义相似性计算领域的相关问题。实验结果表明,所提方法在Rubenstein数据集上与人工判断有着0.910 9的相关性,与其他经典的相似性计算方法相比有着更高的准确性。

The traditional distance-based similarity calculation method was described.Concerning that the method of distance calculation does not contain sufficient semantic information,this paper proposed an improved method which used WordNet and edge weighting information between the concepts to measure the similarity.It considered the level of depth and density of concepts in corpus,i.e.the semantic richness of concept.Using this method,the authors can solve the semantic similarity calculation issues and make the calculation of similarity among concepts easy.The experimental results show that,the proposed method has a 0.910 9 correlation with the benchmark data set-Rubenstein concept pairs.Compared with the classical method,the proposed method has higher accuracy.