MAFAWD: An Adaptive Weight Distribution Clustering Algorithm Based on Multi-layer Attribute Fusion

With the rapid development of social networks, the structure of the data presents more and more complex, and the graph data appears now. This paper proposes an adaptive clustering algorithm based on Multi-layer Attribute Fusion for the Adaptive Weight Distribution (MAFAWD) algorithm to solve the graph data clustering problem. It combines the graph node attribute and the structure relation through the pre-set merger. In this paper, we use some rules to unify the graph node attribute and structure into the same network for clustering, and the influence of node attribute and structure relation on clustering results is considered synthetically. In the multi-layer attribute fusion model, we divide attribute layer and structure layer. Considering the different clustering influence of the graph node attribute and structure, we set different weight layer coefficients. This paper uses affine propagation clustering algorithm and node voting mechanism to change the different weight layer coefficients. It makes the data reflects the original distribution and has a better clustering result. Finally, we verify the algorithm on the real DBLP data set. The experimental result on real data set shows that the MAFAWD algorithm has a better clustering result through compared with the traditional graph clustering algorithms.