A fast connected component algorithm based on hub contraction

Finding the connected components of an undirected network is a fundamental computational problem at the heart of many network applications and has been applied in many fields, such as communication network robustness and computer vision. This paper presents a new form of graph contraction referred to as the “hub contraction,” in which the node with the maximum degree is selected as the central node, after which this node and its immediate neighbors are combined to form a new central node. Based on the hub contraction principle, we propose a fast connected component algorithm, which exploits the existence of high-degree nodes to obtain the number and size of connected components. The efficiency of the proposed algorithm was verified by comparing it with other connected component algorithms within random networks and scale-free networks. The results showed the performance of our algorithm to be superior in terms of the processing time it requires, most notably when applied to scale-free networks.