Research On Overlapping Community Discovery Algorithms Based On Clustering Integration

There are many kinds of complex networks in real life.With the development of the times,more and more complex networks have been studied in depth.Overlapping community detection algorithms,as a tool that can help people understand complex networks,play an important role in daily life.Since the phenomenon of overlapping communities in real networks is more common,the research on overlapping communities detection algorithms is of great significance.Although the current research on overlapping community detection algorithms has achieved great success in complex networks,there are still some problems.Existing overlapping community detection algorithms mostly focus on a single aspect of research,which will make some nodes unable to be accurately divided into the communities they belong to when the information in the network is not sufficient,that is,division errors.There are also some overlapping community detection methods that do not take into account the influence of nodes on the entire network,which makes the division results somewhat biased.In this paper,an in-depth study of the inaccurate community division in existing overlapping community detection algorithms is carried out.The specific research content is as follows.Firstly,in order to dig deeper into the influence of nodes on the entire complex network,an overlapping community detection algorithm based on feature extraction is proposed.The algorithm integrates multiple non-overlapping community detection algorithms to obtain a set of multiple community division results,and retains the diversity of the division results.At the same time,the division result is combined with the membership metric to generate a low-dimensional vector representation of each node,and the vector of the node is learned to discover overlapping communities in complex networks.Secondly,to solve the problem of inaccuracy and instability of a single division method,an overlapping community detection algorithm based on similarity and modularity is proposed.The algorithm also integrates multiple non-overlapping community detection algorithms to preserve the diversity of the division results.At the same time,the community similarity and local modularity are combined to screen the initial community division results,leaving high-quality communities in the set of candidate communities.Achieving the best of the same category has effectively improved the effect of community detection.Finally,for the algorithm proposed in this paper,experiments were carried out on real data sets and synthetic data sets,and the results of overlapping communities were compared with the existing experimental results of overlapping communities,which verified the effectiveness of the algorithm in this paper.