Partitioning Networks with Node Attributes by Compressing Information Flow

• DOI: 10.48550/arxiv.1405.4332
• Publication Date: 2014-01-01
• Authors: Laura M. Smith; Linhong Zhu; Kristina Lerman; Allon G. Percus
• Keywords: Information flow; Network partition

Real-world networks are often organized as modules or communities of similar nodes that serve functional units. These also rich in content, with having distinguishing features attributes. In order to discover a network's modular structure, it is necessary take into account not only its links but node We describe an information-theoretic method identifies by compressing descriptions information flow on network. Our formulation introduces content the description flow, which we then minimize groups attributes tend trap information. The has several advantages: conceptually simple and does require ad-hoc parameters specify number control relative contribution network structure. apply proposed partition real-world known community demonstrate adding helps recover underlying structure content-rich more effectively than using alone. addition, show our faster accurate alternative state-of-the-art algorithms.