An Approach to Analyze the Vulnerability of Function-Based Social Networks Using Clustering Coefficient

Robustness in response to unexpected events is always ideal for real-world networks. In order to improve the robustness of any network system, it is important to analyze the vulnerability to external interference such as accidental fault or attacks that are introduced into the network elements. In this paper, a novel study investigates the robustness of complex network using the clustering coefficient of networks against loss of elements. In particular, we can identify vertexes that can damage the network due to the weakening of its clustering, which is evaluated by means of the average of clustering coefficient. This is important because any tangible change that leads to a clustering is caused by defects that can reduce network functionality, such as the ability to spread information on a social network. We present this risk analysis as an optimization method and demonstrate the completeness and uncertainty of our approach to identify main vertices for clustering. Finally, we perform comprehensive experiments in the combined social networks that are generated by different models. The experimental results show that the average clustering coefficient is very efficient in analyzing the fault of network node failure. Also, the results confirm that the technique of removing the important nodes, especially in terms of the closeness centrality, is very effective in the analysis of clustering vulnerability.