基于关键词的关系数据库查询技术研究

项目名称： 基于关键词的关系数据库查询技术研究

项目编号： No.61303004

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 林子雨

作者单位： 厦门大学

项目金额： 23万元

中文摘要： 采用关键词对关系数据库进行查询，是当前数据库领域的热点研究问题，它不需要用户了解数据库模式知识，也不需要用户掌握专业的结构化查询语言（比如SQL语句），用户只要使用关键词即可查询数据库内容，具有很好的应用前景。目前，对于基于关键词的关系数据库查询问题，学术界在以下三个方面的关键技术研究还不够深入：（a）如何支持高效的top-k聚合关键词查询；（b）如何尽量降低单个查询的代价；（c）如何根据不断变化的用户兴趣动态优化查询结果。本课题将针对上述三个挑战展开相关研究，拟达到以下目标：（1）提出嵌入到RDBMS的、基于聚合单元覆盖的top-k聚合关键词查询方法；（2）提出基于数据图和"渐进"蚁群优化算法的关键词查询方法，有效降低单个查询代价；（3）提出基于概念漂移理论和兴趣节点集的动态查询结果优化方法；（4）将构建原型系统，并发表10篇以上论文。

中文关键词： 关系数据库；关键词查询；数据图；结构化查询语言；蚁群优化算法

英文摘要： Keyword search over relational databases has emerged to be a critical problem. It makes it an easier task to retrieve information from structural data, because users, especially casual users, do not need to learn query languages such as SQL or to know anything about data schemas. However, the existing work is not enough on the following three aspects: (a)how to support efficient top-k aggregate keyword search; (b)how to reduce the cost for a single search as much as possible, because the keyword search problem is seen as a minimum Steiner-tree problem in those methods based on data graph, and the minimum Steiner tree problem is proved to be an NP-hard problem; and (c) how to dynamically optimize query results according to the dynamically changing user interest. Therefore, our research work focuses on the following three aspects: (1)providing built-in top-k aggregate keyword search capabilities in RDBMS by a new efficient approach based on aggregate unit cover; (2)proposing a novel keyword search method based on progressive ant colony optimization and data graph, so as to reduce the cost of a single search; and (3)proposing a novel method of dynamic optimization of search results based on concept drift and user interest nodes. A prototype will be developed to carry out the experiments. The academic achievements w

英文关键词： RDBMS；keyword search；data graph；SQL；ant colony optimization