An Effective Network Partitioning Algorithm Based on Two-Point Diffusing Strategy

Abstract

The network modeling and analysis have played important roles in fields of physics, sociology, biology, and computer science. Recently, community structure has been considered as an important character for complex networks, and its detection can bring great benefit in real world affairs. In the paper, a new heuristic algorithm based on two-point diffusing strategy is proposed. At first, two pseudo-core points are identified according to the clue of the longest path in a network. Then, two embryonic communities and an undecided node set are generated through performing diffusing operation on such two points. Subsequently, an experience rule is used to classify the undecided nodes to form the final community structure. In addition, the effectiveness and efficiency are validated by comparison experiments with four real-world networks. The experiment results show that our TPD algorithm can yield better community partition results and shorter computing time than the existing classical community detecting algorithms.