基于CPN网络的Deep Web数据语义标注  

Semantic Annotation Based on CPN Network for Deep Web Data

作　　者：马安香[1] 高克宁[1] 张晓红[1] 张斌[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110004

出　　处：《东北大学学报（自然科学版）》2009年第6期794-797,共4页Journal of Northeastern University(Natural Science)

基　　金：国家自然科学基金资助项目(60773218);辽宁省科学技术基金资助项目(20072031)

摘　　要：全面准确地标注Deep Web数据是实现Deep Web数据集成系统的关键问题,然而现有的DeepWeb数据语义标注方法还不能很好地解决这一问题.提出一种基于CPN网络的Deep Web数据语义标注方法,通过提取属性值的基本特征,采用CPN网络实现Deep Web数据语义标注.同时,采取了一种有效的方法准确获取Deep Web结果页面中的属性值,为语义标注奠定了良好的基础.与同类成果相比,基于CPN网络的Deep Web数据语义标注方法提高了语义标注的准确率及召回率.An overall and accurate annotation of Deep Web data is the key to Deep Web data integration system. However, the existing methods of semantic annotation are unavailable to solve the problem well. An approach to semantic annotation based on CPN network is proposed for Deep Web data. By extracting the basic features of attribute values, semantic annotation is implemented with CPN network. At the same time, an effective method is proposed to obtain the attribute values on Deep Web result pages, thus laying a sound foundation for semantic annotation. Experimental results showed that the approach to semantic annotation based on CPN network improves the accuracy and recall.

关 键 词：DEEP WEB数据集成 语义标注 CPN网络 特征选取 分隔符序列 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]