Integration of graphs from different data sources using crowdsourcing

详细信息    查看全文

• 作者：Younghoon Kim^a ; Woohwan Jung^b ; Kyuseok Shim ; ^b ; ^{shim@kdd.snu.ac.kr}
• 关键词：Graph integration ; Crowdsourcing ; Entity resolution
• 刊名：Information Sciences
• 出版年：2017
• 出版时间：April 2017
• 年：2017
• 卷：385-386
• 期：Complete
• 页码：438-456
• 全文大小：2732 K
• 卷排序：385

文摘

We develop an algorithm to integrate two graphs collected from different sources. We propose an active graph matching algorithm to exploit crowdsourcing. We suggest a random walk model to compute the matching probability between two nodes. We derive an inference method to estimate the matching probability with EM algorithm. For active learning, we devise a method to select the most informative query node.