Revealing the Hidden Language of Complex Networks

Abstract

Sophisticated methods for analysing complex networks promise to be of great benefit to almost all scientific disciplines, yet they elude us. In this work, we make fundamental methodological advances to rectify this. We discover that the interaction between a small number of roles, played by nodes in a network, can characterize a network's structure and also provide a clear real-world interpretation. Given this insight, we develop a framework for analysing and comparing networks, which outperforms all existing ones. We demonstrate its strength by uncovering novel relationships between seemingly unrelated networks, such as Facebook, metabolic, and protein structure networks. We also use it to track the dynamics of the world trade network, showing that a country's role of a broker between non-trading countries indicates economic prosperity, whereas peripheral roles are associated with poverty. This result, though intuitive, has escaped all existing frameworks. Finally, our approach translates network topology into everyday language, bringing network analysis closer to domain scientists.

From Theory to Practice

“…But new data analytics engines face new kinds of workloads, where multiple large tables are joined, or where the query graph has cycles. For example, Yaveroǧlu et al [37] have recently discovered that the structure of a complex network can be characterized by counting various patterns in the graph. Each pattern, called a graphlet, represents a small graph.…”

Noise-tolerant model selection and parameter estimation for complex networks

“…The ''model fitting'' process aims at finding the best model being capable of generating networks similar to the target network (Model Selection), and then estimating the model parameters which generate the most similar graphs to the target network (Parameter Estimation). Applications of network model fitting include simulation of network dynamics [7][8][9], graph summarization [10,5,[11][12][13], network sampling [14][15][16][17][18], network anonymization [19][20][21][22], and network comparison [23,10,[24][25][26][27]11].…”

Modeling topologically resilient supply chain networks

“…(1), the problem of finding an action matrix M is framed as a multiobjective problem. The decision to frame this within a multiobjective context is based on numerous observations in network science literature arguing that it is a robust approach to determining generator suitability (Pržulj 2007;Harrison et al 2016;Fay et al 2014;Yaveroǧlu et al 2015). To solve this multi-objective search problem, we implement Pareto Simulated Annealing (PSA) (Czyzak and Jaszkiewicz 1998), as it is known to be a useful metaheuristic capable of global optimization in a large search space in a fixed amount of time.…”