中文摘要：

关系数据库上的关键词查询使得用户不必了解SQL语法或者数据库模式即可方便进行检索,并利用关系表的连接来保证查询的完整性。由于关键词集合查询的信息通常与不同数据类型的字段相关,并且分散在多个表中,而关系数据库缺乏对这类信息有效的索引,从而导致查询效率和准确率比较低。针对关系数据库的结构化模式,提出了一种基于层次树的数据提取方法,将关系数据库模型化为一棵层次树,关键词查询问题转化为树的遍历问题,并从关键词数据类型的分类、属性值之间的相关度、查询结果评分排序等方面设计了一个综合的优化方法。通过实验验证,该方法可以实现以较高的效率和较低的计算代价完成关键词查询。

英文摘要：

Keyword search on relational databases facilitates users＇ information retrieval without any knowledge of the database schema or SQL grammar,and it guarantees the completeness of the result by joining tables.Since the data needed is usually related to many different data types and scattered in multiple tables,and relational databases lack the effective indexes of such information,this kind of keyword search is in low efficiency and accuracy.According to the structured schema of relational databases,an approach is proposed for retrieving structured data based on hierarchy tree,modeling the relational database as a hierarchy tree,under this theoretical framework,keyword searching is transformed into a traversing problem in the tree.The comprehensive methods is designed based on keyword classification,attribute values dependency,and results ranking to guide the traversing process.The experimental results show the high effectiveness and efficiency of proposed solutions.