【24h】

A Graphical Method for Reference Reconciliation

机译:参考核对的图形方法

获取原文

摘要

In many applications several references may refer to one real entity, the task of reference reconciliation is to group those references into several clusters so that each cluster associates with only one real entity. In this paper we propose a new method for reference reconciliation, that is, in addition to the traditional attribute values similarity, we employ the record-level relationships to compute the association similarity values of references in graphs, then we combine this kind of similarity with the traditional attribute values similarity and use the clustering algorithm to group the closest references.
机译:在许多应用程序中,多个引用可能引用一个真实实体,引用协调的任务是将这些引用分组为几个群集,以便每个群集仅与一个真实实体关联。在本文中,我们提出了一种新的引用对帐方法,即,除了传统的属性值相似度之外,我们还使用记录级关系来计算图中引用的关联相似度值,然后将这种相似度与传统属性重视相似性,并使用聚类算法对最接近的参考进行分组。

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号