一种用户主导的跨组织数据按需集成方法  

On-Demand Integration of Cross-Organizational Data in a User-Guided Way

作　　者：温彦[1,2] 刘晨[2] 韩燕波[2] 

机构地区：[1]中国科学院计算技术研究所,100190 [2]北方工业大学云计算研究中心,100144

出　　处：《西安交通大学学报》2013年第2期116-123,共8页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(60903137;61033006)

摘　　要：为了在互联网上实现跨组织数据共享,以及及时响应多变的临机业务的需求,在数据服务模型的基础上,提供了一种以用户为主导的数据服务组合方法。该方法可帮助用户按需集成跨域数据,以基于概率的模式匹配方法作为算子实现的基础,将该方法用于一个交互式的组合环境中,通过向用户不断推荐合适操作来完成组合过程。利用对用户反馈的学习来优化推荐内容和组合过程,通过案例、原型系统和相关工作的比较分析,验证了该方法具有较好的实用性。To enable cross-organizational data sharing on the Internet,and to achieve quick adaptation to situational and changing requirements,a user-guided data service composition approach is proposed based on the data service model.The proposed approach can help users integrate the cross-organizational data in an on-demand manner.A set of data service composition operators is proposed,and uses the probability-based schema matching techniques as its implementation basis.An interactive composing environment is developed,in which the schema matching techniques are used to recommend appropriate operations and users＇ feedbacks are learned to optimize the composition process.The usability of this approach is validated via a use case study,a preliminary implementation and the related work comparisons.

关 键 词：数据服务 按需集成 服务组合 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]