Link assessment in online social networks

The information in this page is based on our work: Are We Really Friends?: Link Assessment in Social Networks Using Multiple Associated Interaction Networks

Link assessment is the process of identifying and eliminating the noise from network systems in order to better understand these systems.

Network with false positive and false negative links
Network with assessed links. Green are true positives, red are false positives, and dashed red are the false negatives.

 

For the above network, red links are false positive links (low intensity links, or just noise) and the dashed links are false negative links (links that should be in the network based on other observation).

 

The Venn diagram for edge overlapping between the Facebook social network SN and the other associated interaction networks G for the research group data set
The Venn diagram for edge overlapping
between the Facebook social network SN and the other
associated interaction networks G for the research group
data set

We employed a machine learning classifier for assessing the links in the social network of interest using the data from the associated interaction networks around it, e.g, any auxiliary information other than the social network itself.

Link assessment machine learning model
Link assessment machine learning model

Two data sets were used to test the proposed method, and the results showed that it is possible to assess the links in a social network using any network in the associated interaction networks. Also, the results revealed that the assessment using the interaction networks is slightly better than the assessment using the social network itself, which suggest the existence of a correlation between the associated interaction networks and the formation process within the social network.