摘要
模式匹配在诸如数据集成、数据仓库、信息共享和计算机网络交换等许多应用领域起到关键作用。目前,自动模式匹配方法也不能解决复杂模式情况下的匹配问题。本文提出一种基于关系模式领域中实体分类的数据库模式匹配方法。该方法通过朴素贝叶斯学习将实体分为不同的类(子模式),然后以同样的类来匹配子模式之间的模式元素。本方法在复杂模式条件下可有效提高匹配效率,减少匹配工作量,节省人力资源。
Schema matching plays a key role in many application domains,such as data integration,data warehouse, and information share and exchange on computer network. Currently,approaches of automatic schema matching cannot solve matching issue under the circumstance of complex schema well. This paper introduces an approach based on entity classification in the domain of relation schema. It divides entities into different categories (sub-schema) using Na ive Bayes Learning ,and then matches schema elements between the sub-schemas with the same category. It can effectively improve matching results, reduce the number of element-to-element comparisons and save user efforts under the circumstance of complex schema.
出处
《计算机科学》
CSCD
北大核心
2004年第10期157-159,F004,共4页
Computer Science
基金
国家"973"重点基础研究发展规划项目(G1999032705)
国家"十五"科技攻关计划(2001BA102A01)