Ensembles of graphs arise in several natural applications, such as mobility tracking, computational biology, socialnetworks, and epidemiology. A common problem addressed by many existing mining techniques is to identify subgraphs of interest in these ensembles. In contrast, in this paper, we propose to quickly discover maximally variable regions of the graphs, i.e., sets of nodes that induce very different subgraphs across the ensemble. We first develop two intuitive and novel definitions of such node sets, which we then show can be efficiently enumerated using a level-wise algorithm. Finally, using extensive experiments on multiple real datasets, we show how these sets capture the main structural variations of the given set of networks and also provide us with interesting and relevant insights about these datasets.
- Date of publication:
- November 14, 2015
- IEEE International Conference on Data Mining