Research On Mining To Hierarchical And Overlaying Community Structure In Complex Networks

The present society is an era of big data,data mining is a new kind of technology for large-scale data processing methods.It can help users find hidden rules and patterns in large amounts of data.Complex networks,such as social networks,can find hidden rules and patterns,faces a big challenge because of its great complexity.Complex networks are usually present a community structure characteristics,how to efficiently find the community structure feature of complex network is currently one of the hot spot of the research on complex networks.Current research in the field has a lot of excellent community mining algorithms,but these algorithms are better only to find out the hierarchy or overlapping.Can also reflect the hierarchy and overlapping community structure of complex networks for the accurate dig out the potential information of the network is very important,in this paper,the research of complex network community structure hierarchy and overlaps the algorithm reveals,the main work is as follows:(1)Existing network hierarchy BGLL algorithm can efficiently and accurately dug up hierarchical community structure of complex networks.But the reality of the network is both gradation and overlaps,in order to more accurately dug up a complex network of potential information,this paper restore the existing excellent network hierarchy community mining algorithm BGLL i n the related software.Then we improved BGLL algorithm,making it not only can find the network hierarchy but also the bridge nodes between two commun ities based on the basic concept of the degree of contribution which module node joining clubs.We optimized the algorithm so that it reflects the hierarchy and overlapping of the complex networks.(2)To test BGLL to improve the performance of the algorithm proposed in this paper,based on related software restore two classic community mining methods,including GN algorithm based on the number of side-mediated and Fast Newman algorithm.And we compare improved BGLL algorithm with classical algorithm and LFM algorithm which reflect the hierarchy and overlapping at the same time with different networks.Comparing the performance indicators of the four algorithms.Application related software tools,this paper implements the BGLL improved algorithm is proposed,and the simulation experiment was carried out.The improved algorithm BGLL simulation results show that the improved algorithm can improve the testing benchmark figure network calculate module degrees than BGLL algorithm,and other mining algorithms improve numerical mining community structure shows that the improved algorithm is more obvious,the effect is much better.At the same time,the improved algorithm c an find the overlapping points in benchmark figure,mining results is better than of LFM algorithm.This further shows that the improved BGLL can reveal more accurately to the hierarchy and overlapping community structure of complex networks.