摘要
区分相同属性是异构数据库环境下语义集成中的一个重要环节,主要的方法是用特征描述属性来评估属性之间的相似性。虽然这种方法具有较高自动化和易于实现的特点,但它将花费更多的时间来比较所有的属性且不能在语义集成中实现并行计算。本文提出了一种基于数据类型的方法来实现异构数据环境下相同属性的确定,这种方法具有在描述比较时间的同时实现语义集成的并行计算的特点。实验结果表明我们的方法能提高系统性能并且不降低查准率和查全率。
Identifying corresponding attributes is an important issue of semantic integration in heterogeneous databases. The main method uses the characteristics describing attributes to evaluate the similarity of attributes. Although this method has highly automated and easily realized characteristics, it will cost more time to compare all attributes, and cannot realize the parallel computation of semantic integration. Accordingly, in this paper we present a data-type-based approach to identify corresponding attributes in heterogeneous databases. Our approach has the characteristics of decreasing the comparing times as well as realizing the parallel computation of semantic integration. The experimental results show our method can improve the system performance without reducing the precision ratio and recall ratio.
出处
《计算机科学》
CSCD
北大核心
2004年第9期96-99,共4页
Computer Science
基金
国家自然科学基金(No.60073047)