Clustering Coefficient Increasing Local Link Switching Algorithm

Abstract

This paper proposed a local link switching algorithm which effectively increases the clustering coefficient of a network while preserving the network node degree distributions. This link switching algorithm is based on local neighborhood information. Link switching algorithm is widely used in producing similar networks with the same degree distribution, that is, it is used in ‘sampling’ networks from the same network pool. Therefore, the switching pairs of links are selected rather ‘globally’ from the network. The proposed algorithm focus on increasing an important network characteristic while selecting candidate pairs of links locally. Clustering coefficient characterizes the relative tightness of a network and is a defining network statistics that appears in many ‘real-world’ network data. Simulation results on three different types of model networks have demonstrate the effectiveness of this algorithm.