A new technique of selecting an optimal blocking method for better record linkage

作者:

Highlights:

• Current evaluation metrics fail to select optimal blocking methods for record linkage.

• Proposed evaluation technique allows for better comparison between blocking methods.

• Proposed selection technique selects top performing blocking methods for new datasets.

• Experimental evaluations demonstrate the effectiveness of these novel techniques.

摘要

•Current evaluation metrics fail to select optimal blocking methods for record linkage.•Proposed evaluation technique allows for better comparison between blocking methods.•Proposed selection technique selects top performing blocking methods for new datasets.•Experimental evaluations demonstrate the effectiveness of these novel techniques.

论文关键词:Record linkage,Blocking,Entity resolution

论文评审过程:Received 13 March 2018, Revised 15 June 2018, Accepted 16 June 2018, Available online 22 June 2018, Version of Record 1 July 2018.

论文官网地址:https://doi.org/10.1016/j.is.2018.06.006