Structural Balance to model trust and mistrust in social networks
This article talks about the polarizing trends of relationships with both trust and mistrust connections in a social networks. They first analyze relationships that are not strongly connected. When there is a subset of the network that is strongly connected containing mistrust, the effect on opinions to the rest of the network neutralizes to 0 when then subnetwork does not satisfy the structural balance theory. However, when the strongly connected network containing mistrust satisfies the structural balance theory, the network polarizes and the the nodes split between the two opinions of trust and mistrust. This article uses the DeGroot model, which analyzes how each node overrides their opinions to the average of its neighboring relationships.
This connects to what we talk about in class regarding the structural balance theory, as we discussed the balance theorem as if a signed complete graph has all positive edges, or if the nodes can be split into 2 sets X and Y, where X and Y both contain all positive edges, and all edges between X and Y are negative. In correlation, in the trust mistrust model, the strongly connected network containing mistrust which satisfies structural balance theory eventually polarize into two sides of opposite opinions, as we discussed in class. Based on the average of the neighboring agents, the rest of the network’s opinions are influenced and end up lying in between these two opposite sides. This was a very interesting phenomenon to learn about, as when thinking about forming opinions on others, the Degroot model seems very intuitive as we often see others that we are not “strongly tied” with through the opinions of our strongly tied peers.