Graph summarization for attributed graphs

The increasing popularity of graph data in various domains has led to a renewed interest in understanding hidden relationships between nodes in a single large graph. And graph summarization is to find a concise but meaningful representation of a given graph. In this paper, we studied the problem of summarizing graph with content associated with nodes. We propose a graph summarization algorithm AGSUMMARY, which achieves a combination of topological and attribute similarities. Our method utilizes the Minimum Description Length (MDL) principle to model the graph summarization problem into a code cost function, and compute an optimal summary of graph with neighborhood greedy strategy. It requires no specified number of summary parts and its running time scales linearly with graph size and the average degree of nodes. Experimental results demonstrate the effectiveness and efficiency of our proposed method. For further illustration, a case example is given to explain our method.