语义Web中的概念等级匹配算法  

A Matching Algorithm in Semantic Web

作　　者：尹东斐[1] 梁力[1] 

机构地区：[1]西安交通大学电子与信息工程学院,西安710049

出　　处：《西安交通大学学报》2009年第4期49-51,74,共4页Journal of Xi'an Jiaotong University

摘　　要：针对语义Web匹配方法仅适于字符串、字典近义词匹配而导致精度低、效果差的问题,提出了一种基于概念等级的语义Web匹配算法.对于一个包含大量概念的本体,算法可根据概念之间的联系建立起概念结构,通过计算获得本体概念之间的相互支持度,从而使得普通概念的支持度低而特殊概念的支持度高,进而将概念的支持度量化为概念等级.以计算得到的概念等级为权值,将概念间的语言学匹配度加权,由此计算出新的概念的匹配度.该匹配度可将各个概念之间的内在联系关联起来,从而提高语义匹配的精度.实验结果表明,所提算法在本体间的语义匹配精度比字典近义词匹配法提高了20%.A conception rank based algorithm for semantic Web is presented to improve the low efficiency and poor precision of the synonymous words matching algorithm and the string matching algorithm. For an ontology with large amount of conception, the conception structure of the ontology is constructed by the algorithm based on the relationship among conceptions. Then the algorithm calculates the contribution for each conception such that the general conception has low contribution and the special conception has high contribution, and then the contributions are quantified into conception ranks. The resulting conception ranks are used as weights and the weighted linguistically matching degree is used to calculate a new conception matching degree in semantic matching. The new matching degree can improve the accuracy of semantic matching because the structure coefficient is taken into account. Experiments show that the matching degree can be increased by 20% compared with the synonymous words matching algorithm.

关 键 词：语义WEB 概念等级 概念结构 语义匹配 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]