Abstract

Previous chapters focus on the study of static interactions, represented by a binary number or a positive weight. Nevertheless, in many application domains, interactions vary over time. The longitudinal nature of such datasets calls for replacing classical graph-based models with temporal network models represented by tensors (Holme and Saramäki, 2012; Kivelä et al., 2014). We note that taking into account the temporal aspects not carefully, e.g., by aggregating or smoothing the temporal data along the time axis, can lead to a loss of valuable information.