基于本体的Deep Web自动标注方法研究

Research on Ontology-based Automatic Annotation for Deep Web

结合基于查询接口模式的Deep Web标注方法,提出一种基于网页视觉信息的Deep Web标注方法,用本体词组去替换原有标注信息,这种替换确保标注信息的一致性,可以很好地弥补原有方法的许多缺陷,并且有效提高原有方法的准确率和召回率。

This paper puts forward a kind of labelling method according to the Web page sense of vision information, which is also based on the inquiry connection pattern labelling method. Then it uses ontology phrases to replace the original label information, and the replacement of label information can ensure the labelling information uniformity. This method is a good way to make up the defects of original method, and effectively improves the precision and recall.