基于改进相似度传播算法的本体匹配

摘要

为解决本体异构、实现不同本体应用程序间互操作以及数据集成,提出一种基于RDF图的改进相似度传播匹配算法.首先通过WordNet发现初始相似对种子,经过预处理把本体表示成RDF三元组形式,针对RDF图的特点,将相似度传播的条件扩展到三元组中,发现可能相似对；然后采用综合元素特征的方法计算相似度.相似度传播、发现可能相似对种子、相似度计算是一个循环迭代的过程,直到满足收敛条件.实验表明了该算法的有效性,并在时间性能上也有所提高.%In order to solve the semantic heterogeneity and achieve interoperability between Web applications of different ontology and integrating data, an improved matching algorithm of similarity propagation based on RDF graph was proposed. First, it sought to find initial similar seeds by WordNet. Then it expressed ontology as RDF triples through preprocessing. According to the characteristics of RDF graph, it expanded similarity propagation to triples to find probable similar pairs and then calculated similarities by elements' features. The procedure of similarity propagation, finding probable similar pairs and calculating similarities is a cyclic iterative process until it is convergent. The experimental results show that the algorithm is effective and has better time performance.