A Multilayer Graph Approach for Predicting Computer Network Cyber-attacks

Today's society is heavily oriented towards digitalization, which increasingly affects the management of cities and services. This process is performed through the use of the Internet of Things (IoT) paradigm, from which arise problems related to security. In this scenario, based on the continuous exchange of information on the network, an increasingly significant role is played by systems able to guarantee data security. Protecting the modern Computer Networks could be a very complex task. In this paper, a methodology based on three graphic models (Context Dimension Tree, Ontology and Bayesian Network) is proposed. Three different models are used which use context representation and probabilistic approaches to predict cyber-attacks. The paper proposes, in fact, the use of Bayesian networks built through an ontological definition of the problem dropped on a certain context represented by a Context Dimension Tree. The proposed approach has been experimented in a real scenario providing satisfactory results.