Research on Community Discovery Method of Complex Network Based on Density Peak Clustering and Its Application

Nowadays, examples of complex networks are everywhere, such as biological networks, transportation networks and power networks. Community discovery is a basic task of complex network analysis. In complex network analysis, analyzing community structure and extracting information are crucial research tasks. The existing research starts with the density peak clustering algorithm and obtains a variety of methods to divide complex network communities, but the sensitivity of cutoff distance d
c
is not considered in the research of density peak clustering algorithm. Based on this, this paper proposes a new method of density peak clustering in complex network communities (GWO-DPC-Node2vec) based on grey wolf optimization. Firstly, the grey wolf optimization algorithm is iterated to find the optimal d
c
value of the density peak species silhouette coefficient. Compared with the classical algorithm, the stability and superiority of the improved density peak clustering algorithm are proved. Then, the nodes in the network are transformed into vectors through Node2vec, and the grey wolf optimization algorithm is run to search the cutoff distance with the maximization of Q value of modularity as the optimization goal. The searched d
c
results are applied to the density peak clustering algorithm, so that the division results can be obtained. Finally, the stability and superiority of the complex network community discovery method proposed in this paper are demonstrated by real network data sets.