Link Modeling in Interaction Networks

Links in interaction networks are results of established communications or agreements between autonomous agents. Links are created upon will of agents to make communications with each other. In this view, links are not the associations subjectively assumed between objects. Moreover, links are not necessarily constant over time. Modeling this type of links in social networks helps to detect communities, predict friendships and undetstand and control of rumor spread.

Simulated Networks

While real world data can be obtained abundantly from social networks, but, simulated networks are more suitable for theoretical studies, because underlying assumptions about their structures are not doubtful.

Barabasi-Albert Model

Here, 100 networks, each one with 10,000 nodes have been generated based on the scale-free model of Barabasi-Albert.

Centrality indeces

For every node of each network, four centrality indeces have been computed as follows,

• Ce — Eigen-vector centrality,
• Ck — Katz centrality,
• Cc — Closeness centrality,
• Cb — Betweenness centrality.

download

Contact

Ali Katanforoush, <a_katanforosh@sbu.ac.ir>, Department of Computer Science, Shahid Beheshti University, G.C., Tehran.