Research And Application Of Overlapping Community Detection Algorithms In Complex Networks

The formation of complex systems is closely related to the development of society and the progress of science and technology,and the relationship between individuals is the key to the formation of complex systems.It is also very difficult to extract valuable information from a variety of complex systems.Therefore,as an abstract form of complex systems,complex networks have been proposed by researchers.On this basis,researchers have found the general laws and characteristics of the existence of networks.With the discovery of small world characteristics,scale-free characteristics and community structure characteristics in complex networks,the research of complex networks has reached a climax.Community is a kind of clustering structure,which is composed of nodes and edges formed between nodes.The nodes are closely connected within the community and sparse between the communities,which is the most important feature of the community structure.In the real world,community structure usually exists in an overlapping way,that is,there is intersection between communities,and one node may belong to multiple communities.The detection of overlapping communities is also becoming more and more common in real networks.Community structure plays an important role in information mining in many fields.Therefore,it is of great significance to study the characteristics of community structure in complex networks.This paper first analyzes and studies the complex network,community structure,community detection algorithm,community structure evaluation index,LPA algorithm,COPRA algorithm improvement and application.The COPRA algorithm is widely used because of its simplicity and efficiency.This paper would improve the COPRA algorithm in terms of label initialization and label propagation.The algorithm proposed in this paper improves the concept of node degree critical value d₀ in the label initialization stage,and proposes the use of asynchronous propagation to replace the original synchronous propagation in the label propagation stage to improve the stability and accuracy of the COPRA algorithm.In the experimental part,this paper compares the accuracy and time complexity of COPRA algorithm,CPM algorithm and improved COPRA algorithm in real network datasets and LFR dataset.After comparison of the results,the proposed algorithm can find a higher quality community structure.At the end of paper,the DBLP dataset is introduced.Some data is extracted from the parsed DBLP dataset to construct the author's collaborative network.The algorithm proposed in this paper is used to find the existence of overlapping communities in the dataset and verify the accuracy of the algorithm.The result proves that the algorithm can be applied to a large data set.