Diversified Coherent Core Search on Multi-Layer Graphs

Mining dense subgraphs on multi-layer graphs is aninteresting problem, which has witnessed lots of applications inpractice. To overcome the limitations of the quasi-clique-basedapproach, we propose d-coherent core (d-CC), a new notionof dense subgraph on multi-layer graphs, which has severalelegant properties. We formalize the diversified coherent coresearch (DCCS) problem, which finds k d-CCs that can cover thelargest number of vertices. We propose a greedy algorithm withan approximation ratio of 1 1/e and two search algorithmswith an approximation ratio of 1/4. The experiments verify thatthe search algorithms are faster than the greedy algorithm andproduce comparably good results as the greedy algorithm inpractice. As opposed to the quasi-clique-based approach, ourDCCS algorithms can fast detect larger dense subgraphs thatcover most of the quasi-clique-based results.