Computer Engineering and Applications ›› 2012, Vol. 48 ›› Issue (36): 160-163.

Previous Articles     Next Articles

Research on entity extraction method of Deep Web data integration

ZHAO Haixia1, LI Daoshen1, LIU Yong1, ZHAO Jiacheng2   

  1. 1.Electronic & Information Engineering College, Henan University of Science and Technology, Luoyang, Henan 471003, China
    2.Software College, Changchun University of Science Technology, Changchun 130000, China
  • Online:2012-12-21 Published:2012-12-21

一种Deep Web查询结果的实体抽取方法

赵海霞1,李道申1,刘  勇1,赵嘉诚2   

  1. 1.河南科技大学 电子信息工程学院,河南 洛阳 471003
    2.长春理工大学 软件学院,长春 130000

Abstract: Based on the realization of Deep Web integrated query mechanism, Deep Web information can be obtained from the resulting pages, so how to extract the entity information of Deep Web from the results pages effectively becomes the key of Deep Web data integration. A method that combines the index with the edit similarity methods is proposed, which resolves the problem of data extraction of Deep Web result page. Large experimental results show that this approach is feasible, and can improve the precision and recall of Deep Web data extraction.

Key words: Deep Web, data extraction, Document Object Model(DOM) tree, index, similarity

摘要: Deep Web中蕴含着丰富的高质量的信息,通过Deep Web集成查询接口可以获取到包含这些信息的结果页面,因此,Deep Web查询结果页面的数据抽取成为Deep Web数据集成的关键。提出了将索引方法和编辑相似度相结合的方法,来完成Deep Web查询结果页面的数据抽取工作。大量实验结果表明:该方法是可行的,并且能够提高Deep Web数据实体抽取的准确性和召回率。

关键词: 深度网, 数据抽取, 文件对象模型(DOM)树, 索引, 相似度