Introducing multilayer stream graphs and layer centralities

Abstract

Graphs are commonly used in mathematics to represent some relationships between items. However, as simple objects, they sometimes fail to capture all relevant aspects of real-world data. To address this problem, we generalize them and model interactions over time with multilayer structure. We build and test several centralities to assess the importance of layers of such structures. In order to showcase the relevance of this new model with centralities, we give examples on two large-scale datasets of interactions, involving individuals and flights, and show that we are able to explain subtle behaviour patterns in both cases.