【作者】 程伟； 周龙骧； 林河水； 孙玉芳；

【Author】 CHENG Wei~1 ZHOU Long-Xiang~2 LIN He-Shui~1 SUN Yu-Fang~1 (Institute of Software,Chinese Academy of Sciences,Beijing 100080) 1 (Academy of Mathematics and System Sciences,Chinese Academy of Sciences,Beijing 100080) 2

【机构】 中国科学院软件研究所； 中国科学院数学与系统科学研究院；

【摘要】 模式匹配问题即寻找不同模式元素之间的语义对应关系,在数据仓库、异构数据源集成及语义Web等领域都是非常重要的研究基础。目前模式匹配仍大多主要由人工来完成,因此有很大局限性。提出了一种多策略通用模式匹配架构,可以方便地兼容其他匹配策略。采用了一种基于词语间语义距离的方法来计算其语义相似度;提出了一种基于相似度传播的结构匹配算法,有效地考虑了相邻相似节点间的相互影响。实验结果表明这种匹配方法在处理模式匹配任务时能达到较高的精度。更多还原

【Abstract】 Schema matching,the problem of finding semantic correspondences between elements of two schemas, plays a key role in many applications,such as data warehouse,heterogeneous data sources integration and semantic Web.Currently,schema matching is largely performed manually by domain experts,thus a time-consuming and labor-intensive process.In this paper,we describe a multistrategy schema matching framework,which can combine multiple matching strategies flexibly and its architecture is extensible to new matchers.We adopt an approach based on semantic distances between words to compute their semantic similarity.We propose a structural matching algorithm based on semantic similarity propagation,which considers the effect between neighboring nodes.After describing our approach,we present experiment results on several real-world domains,and show that the approach discovers semantic mappings with a high degree of accuracy.更多还原