Machine Learning-Based Context Space Theory

Situation awareness of human and artificial agents can be improved by the recognition and adequate representation of real-life situations. The lack of easy-understandable, easy-to-use, and effective computational models of situations hindered the adoption and diffusion of situation awareness approaches in modern human-machine systems. Context Space Theory is a context awareness approach that uses a geometric metaphor to provide integrated mechanisms for representing contexts and situations. A drawback of this approach is the expert-based definition of context and situation spaces. This process can be time-consuming and expensive. In this paper, we propose a novel approach, namely Machine Learning-based Context Space Theory, which adopts machine learning techniques and, in particular, decision trees, to semi-automatically define context spaces and situation spaces with a data-driven approach. A case study related to the monitoring and control of the Covid-19 pandemic in Italy is proposed to demonstrate the feasibility and benefits of the proposed approach.