中文摘要：

针对现有Deep Web信息集成系统没有考虑查询接口动态性的特点,造成本地接口与网络接口查询能力不对等的问题,提出一种基于演化版本的Deep Web查询接口维护方法。该方法通过构建本地接口的版本化模型来刻画接口的增量变化,识别变动比较活跃的属性集合;然后采取试探性查询来构建最优查询语句,获取网络接口数据源的变动信息,演化出本地接口的下一个版本,实现对本地查询接口数据源的信息维护的迭代过程。实验结果表明,该方法降低了深网环境变化对Deep Web信息集成带来的影响,确保了Deep Web查询接口的准确率和查全率的稳定性。

英文摘要：

In order to solve the problems existed in the traditional Deep Web information integration system that without con- sidering the dynamic feature of search interface, causing local interface and network interface query ability is not equal. Therefore,this paper proposed a Deep Web search interface maintenance method based on evolution version. In this method, constructing the version models of local search interface was to express the incremental change of it ,and to extract the active attribute set. Next, generating the best query string with the set and probing query was to extract the change content and get the next version of local interface. Finally,it could realize the iterative maintenance of local search interface data source. The experimental results show that this method is able to decrease the impact caused by deep Web network changing, and keep the recall and precision of Deep Web search interface in a stable state.