Which similarity measure to use in network analysis: Impact of sample size on phi correlation coefficient and Ochiai index

作者：

Highlights：

• Studied the impact of sample size on an implicit network inferred using Phi Correlation coefficient and Ochiai coefficient.

• Illustrated using a network of diseases developed using 22.1 Million patient records.

• Found Ochiai coefficient to be less sensitive to the sample size than Phi Correlation coefficient.

• The betweenness centrality was most affected by the sample size.

摘要

•Studied the impact of sample size on an implicit network inferred using Phi Correlation coefficient and Ochiai coefficient.•Illustrated using a network of diseases developed using 22.1 Million patient records.•Found Ochiai coefficient to be less sensitive to the sample size than Phi Correlation coefficient.•The betweenness centrality was most affected by the sample size.

论文关键词：Implicit network,Analytics,Comorbidity network,Sample size,Inferred network,Ochiai coefficient,Phi correlation coefficient

论文评审过程：Received 31 May 2020, Revised 18 August 2020, Accepted 19 August 2020, Available online 6 September 2020, Version of Record 6 September 2020.

论文官网地址：https://doi.org/10.1016/j.ijinfomgt.2020.102229