Overlapping community detection using seed set expansion

Abstract

Community detection is an important task in network analysis. A community (also referred to as a cluster) is a set of cohesive vertices that have more connections inside the set than outside. In many social and information networks, these communities naturally overlap. For instance, in a social network, each vertex in a graph corresponds to an individual who usually participates in multiple communities. One of the most successful techniques for finding overlapping communities is based on local optimization and expansion of a community metric around a seed set of vertices. In this paper, we propose an efficient overlapping community detection algorithm using a seed set expansion approach. In particular, we develop new seeding strategies for a personalized PageRank scheme that optimizes the conductance community score. The key idea of our algorithm is to find good seeds, and then expand these seed sets using the personalized PageRank clustering procedure. Experimental results show that this seed set expansion approach outperforms other state-of-the-art overlapping community detection methods. We also show that our new seeding strategies are better than previous strategies, and are thus effective in finding good overlapping clusters in a graph.

A Block-Based Edge Partitioning for Random Walks Algorithms over Large Social Graphs

“…In [8] authors state that, due to the heavytailed degree distributions and large clustering coefficients properties in social networks, considering only the direct neighbors of a vertex allows to construct good clusters (communities) with low conductance. In [24] authors improve this method to detect communities over graph, but neither edge partitioning nor workload balancing problem is studied. Moreover, the overlapping communities approach for graph partitioning are not suitable to Pregel-like systems.…”

“…In the Pregel approach, we consider the block as a set of edges which are "close" one to another, and these blocks become the component units of each partition in computation, but also the allocation units for workload over machines. Similar to the methodology adopted in vertex partitioning [8,24], we propose to compute a set of K blocks by exploring the graph. An edge is allocated to a block based on its distance from this block.…”

“…Obviously the result of the partitioning, in term of size-balancing or communication during the computation, is highly dependent on the choice of the seeds. This problem has been studied in literature for instance in [24] to detect communities on graph or in [7] where authors propose and experiment for the precomputation step of their recommendation algorithm several landmark selection strategies.…”

Stochastic Blockmodel with Cluster Overlap, Relevance Selection, and Similarity-Based Smoothing

“…Non-Bayesian methods for overlapping clustering include clique percolation [24], line graph partitioning [2], ego network extraction [5], low-rank non-negative matrix factorization based modeling [33], and seed set expansion [30]. The clique percolation method assumes that a graph consists of overlapping sets of cliques, and considers adjacent cliques as overlapping clusters in the graph.…”

“…For example, Yang et al [33] presented a model-based community detection algorithm using a non-negative matrix factorization. Recently, Whang et al [30] have proposed an efficient overlapping community detection algorithm using seed set expansion. They presented effective seed finding methods and produced a set of small conductance clusters by expanding the seed sets.…”