A community detection algorithm based on multi-domain adaptive spectral clustering

Abstract

At present in most field's data sets, spectral clustering community detection algorithm is difficult to predict the number of clustering problems, this paper proposes a community detection algorithm based on multi-domain adaptive spectral clustering (MDASC). Firstly based on the local node density composition, combined with graph edge-betweenness structural similarity matrix, normalized spectral clustering, got the biggest feature dimensions k value, so as to achieve the purpose of automatic identification number of the cluster, and finally re-use k-means classical clustering algorithm to cluster the feature vector space. Experiments show that compared with the traditional spectral clustering community detection algorithm, MDASC can construct a more efficient similarity matrix, simulation community structure is more close to the real, can adapt to all kinds of field sample data.