Contrast Subgraph Mining from Coherent Cores

Graph paŠern mining methods can extract informative and useful paŠerns from large-scale graphs and capture underlying principles through the overwhelmed information. Contrast analysis serves as a keystone in various €elds and has demonstrated its e‚ectiveness in mining valuable information. However, it has been long overlooked in graph paŠern mining. Œerefore, in this paper, we introduce the concept of contrast subgraph, that is, a subset of nodes that have signi€cantly di‚erent edges or edge weights in two given graphs of the same node set. Œemajor challenge comes from the gap between the contrast and the informativeness. Because of the widely existing noise edges in real-world graphs, the contrast may lead to subgraphs of pure noise. To avoid such meaningless subgraphs, we leverage the similarity as the cornerstone of the contrast. Speci€cally, we €rst identify a coherent core, which is a small subset of nodes with similar edge structures in the two graphs, and then induce contrast subgraphs from the coherent cores. Moreover, we design a general family of coherence and contrast metrics and derive a polynomial-time algorithm to eciently extract contrast subgraphs. Extensive experiments verify the necessity of introducing coherent cores as well as the e‚ectiveness and eciency of our algorithm. Real-world applications demonstrate the tremendous potentials of contrast subgraph mining. ACM Reference format: Jingbo Shang1, Xiyao Shi1, Meng Jiang2, Liyuan Liu1, Timothy HanraŠy3, Jiawei Han1. 2018. Contrast Subgraph Mining from Coherent Cores. In Proceedings of ACM SIGKDD Conference on Knowledge Discovery and Data Mining, London, UK, August 2018 (KDD’18), 9 pages. DOI: 10.475/123 4