Search

Community detection has been extensively studied for various applications. Recent research has started to explore node contents to identify semantically meaningful communities. However, links in real networks typically have semantic descriptions and communities of links can better characterize community behaviors than communities of nodes. The second issue in community finding is that the most existing methods assume network topologies and descriptive contents carry the same or compatible information of node group membership, restricting them to one topic per community, which is generally violated in real networks. The third issue is that the existing methods use top ranked words or phrases to label topics when interpreting communities, which is often inadequate for comprehension. To address these issues altogether, we propose a new Bayesian probabilistic approach for modeling real networks and developing an efficient variational algorithm for model inference. Our new method explores the intrinsic correlation between communities and topics to discover link communities and extract semantically meaningful community summaries at the same time. If desired, it is able to derive more than one topical summary per community to provide rich explanations. We present experimental results to show the effectiveness of our new approach and evaluate the method by a case study. [ABSTRACT FROM AUTHOR]

Due to the demand for performance improvement and the existence of prior information, semi-supervised community detection with pairwise constraints becomes a hot topic. Most existing methods have been successfully encoding the must-link constraints, but neglect the opposite ones, i.e., the cannot-link constraints, which can force the exclusion between nodes. In this paper, we are interested in understanding the role of cannot-link constraints and effectively encoding pairwise constraints. Towards these goals, we define an integral generative process jointly considering the network topology, must-link and cannot-link constraints. We propose to characterize this process as a Multi-variance Mixed Gaussian Generative (MMGG) Model to address diverse degrees of confidences that exist in network topology and pairwise constraints and formulate it as a weighted nonnegative matrix factorization problem. The experiments on artificial and real-world networks not only illustrate the superiority of our proposed MMGG, but also, most importantly, reveal the roles of pairwise constraints. That is, though the must-link is more important than cannot-link when either of them is available, both must-link and cannot-link are equally important when both of them are available. To the best of our knowledge, this is the first work on discovering and exploring the importance of cannot-link constraints in semi-supervised community detection. [ABSTRACT FROM AUTHOR]

Discovery of communities in complex networks is a fundamental data analysis problem with applications in various domains. While most of the existing approaches have focused on discovering communities of nodes, recent studies have shown the advantages and uses of link community discovery in networks. Generative models provide a promising class of techniques for the identification of modular structures in networks, but most generative models mainly focus on the detection of node communities rather than link communities. In this work, we propose a generative model, which is based on the importance of each node when forming links in each community, to describe the structure of link communities. We proceed to fit the model parameters by taking it as an optimization problem, and solve it using nonnegative matrix factorization. Thereafter, in order to automatically determine the number of communities, we extend the above method by introducing a strategy of iterative bipartition. This extended method not only finds the number of communities all by itself, but also obtains high efficiency, and thus it is more suitable to deal with large and unexplored real networks. We test this approach on both synthetic benchmarks and real-world networks including an application on a large biological network, and compare it with two highly related methods. Results demonstrate the superior performance of our approach over competing methods for the detection of link communities. [ABSTRACT FROM AUTHOR]