Network theory

About: Network theory is a research topic. Over the lifetime, 2257 publications have been published within this topic receiving 109864 citations.

Topology structure and centrality in a java source code

Abstract: Call graph plays a very important role in program analysis, and it has been widely used in software engineering (e.g., program understanding, compiling optimization, regression test, etc). In the present paper, we studied the source code structure of a Java program from a call graph perspective. We check the basal network properties of the call graph. By comparing several familiar centrality measures (betweenness, closeness, eccentricity, HITS-Authority and PageRank) and applying them to the call graph, we also give the centrality analysis of the call graph.

Applications of eigenvector centrality to small social networks

Abstract: This article investigates conceptual and methodological questions that may arise in applying eigenvector centrality to small social networks such as school classes. The focus on small networks brings out surprising and subtle properties related to the interpretation of the measure. We investigate examples where the weighted adjacency matrix of the underlying social network quantifies inter-individual preferences of whom to work with or play with. We show that mathematical operations such as transposition and symmetrization of the weighted adjacency matrix enhances the power of the measure. It is demonstrated that it is not sufficient to work with the original weight matrix. By working with a symmetrized or a semi-symmetrized or a transposed weight matrix different characteristics of the social interaction are revealed. The method chosen depends on the purpose of the investigation. Identifying isolated or popular individuals in the network are also facilitated using these operations.

Dual-system methods in dynamical analysis. Part 1: Variational principles and their application to nonlinear-network theory

Abstract: The feedback connection of a dynamical system and its dual, through nonlinear coupling elements, is shown to generate a pair of fundamental stationary principles. These provide a unified basis for both the classical Lagrangian and Hamiltonian treatments of dynamical systems aud modern work on nonlinear network theory and linear optimal-control theory.

The Evolution and Transformation of Networks: A Study of Private Health Insurance in Ireland

Abstract: The focus of this research is the process of network evolution and transformation in health insurance. It uses network theory to express the dynamics of change in the private health insurance market and it emphasises the structural features of density and sparseness, as well as relational features of strength and weakness of the ties. The study found that the network was characterised by high centrality, high density and weak ties, and cost efficiency was a key outcome. The study concludes that there needs to be an optimal portfolio of both strong and weak ties, in that together they enable information exchange, knowledge-building and innovation.

Exploring the Application of Network Analytics in Characterizing a Conceptual Design Space

Abstract: The ability to effectively analyse design concepts is essential for making early stage design decisions. Human evaluations, the most common assessment method, describe individual design concepts on a variety of ideation metrics. However, this approach falls short in creating a holistic representation of the design space as a whole that informs the underlying relations between concepts. Motivated by this shortcoming, this work leverages network theory to visualize and characterize features of a conceptual design space. To illustrate the utility of network theory for these purposes, a network composed of a corpus of solutions to a design problem and their semantic similarity is derived, and its design properties (e.g., uniqueness and innovation potential) are studied. This network-based approach not only characterizes features of individual designs themselves, but also uncovers more nuanced properties of the design space through studying emerging clusters of concepts. Overall, this work expands on developing research in design, demonstrating the value in applying network analytics to a conceptual design space as an engineering support tool to aid design decision-making.