Information Diffusion in Multilayer Complex Networks

Abstract

Complex networks intricate the connectivity patterns of interactions among numerous inter- connected elements or entities forming the natural systems. All the real world systems consists of systems with multiple subsystems. Such networks consists of entities that share more than one kind of relation between them, which may be symmetric or asymmetric. To effectively represent such systems, multilayer networks are most suitable, where each layer represents par- ticular kind of relationship and nodes represents entities.This thesis focuses on proposing a model for reshaping the dynamics and analysis of information diffusion in multilayer networks, with a particular emphasis on the improvement of information spread. The research focuses on capturing the interplay between multiple information sources in a multilayer network and the development of an efficient model for managing and leveraging the dynamics of information spread in various contexts of social media. newlineThe study and formulation of centrality metric for multilayer networks laying the ground- work for understanding the real world systems and their limitations. Certain nodes in a network may have a more significant impact on information diffusion than others. These nodes, known as the influential nodes, accelerate or hinder the spread of information. Traditional centrality metrics like PageRank lack the ability to capture inter-dependencies and interactions across different layers. The state-of-the-art also lacks the metric to identify such influencers in the context of multilayer networks. The foremost contribution of this research progress towards the formulation of novel centrality metrics (m-PageRank) for multilayer networks lays the groundwork for identifying influencers of multilayer networks. The integrated analysis enables a comprehensive understanding of the nodes roles and importance within the multilayer net- work based on their structural connectivity and influence.The next contribution addresses two relevant aspects prevailing in the domain...