Model And Algorithm For Searching Community Structure Based On Clique Definition

Many complex systems in nature and society can be described by complex networks,and the research on community recognition of complex networks has attracted extensive attention from scholars from all walks of life in recent years,becoming a hot research field.In the actual networks,there are almost no completely independent community structures.Most of networks are composed of many overlapping and interrelated communities.Therefore,for complex networks with overlapping structures,it is of great application value and practical significance to study community identification models and algorithms.The current researches on overlapping communities mainly focus on the improvement of the clique filtering algorithm,but this algorithm has a defect that it cannot divide the individual nodes outside the cliques.On the other hand,the definition of clique is limited to textual descriptions,and there is no mathematical way to express the division of communities based on the definition of clique.In order to solve the above problems,this paper builds an integer linear programming model based on the definition of clique,and formulates the definition of clique sothat the divided community structures have common structural characteristics.And a community division algorithm based on clique definition is proposed,which improves the defect of the clique filtering algorithm so that all nodes in the network can be effectively divided.In the experimental stage,we firstly constructed a 11-node artificial network,and then selected two real networks,including disease network with different traits and protein interaction network.Finally,the algorithm in this paper is compared with the clique filtering algorithm.The experimental results show that the algorithm proposed in this article not only has a good division effect,but also can complement the defects of the clique filtering algorithm.