Lecture
LCN2 seminar April 2024
- Date
- Friday 26 April 2024
- Time
- Location
- Gorlaeus
- Room
- BM. 2.26 (second floor)
65th LCN2 seminar
Speaker: Alberto Ceria (LIACS)
Title: Characterisation and modelling of temporal higher-order networks
Abstract:
Human social interactions are typically recorded as time-specific dyadic interactions, and represented as evolving (temporal) networks, where links are activated/deactivated over time. However, individuals can interact in groups composed of more than two people. Such group interactions can be represented as higher-order events (or hyperlink activations) in a temporal network but are usually overlooked by traditional pairwise temporal network studies.In this talk, I will present methods to characterise and modelling temporal higher-order networks.In the first part, l will introduce methods to analyse jointly topological and temporal properties of higher-order events to identify similarities and differences among networks. We apply these methods to two classes of real temporal higher-order networks (physical contacts and collaboration networks). We show that, in physical contacts, events close in time tend to occur also close in topology, while this is almost absent in collaboration networks. This result is also supported by the temporal correlation of local events observed in physical contacts, but negligible in collaboration networks. We further observe that, in both classes of networks, nodes involved in many events of a given order are likely active in events of another order.In the second part of the presentation, I will present a memory-based model to predict temporal higher-order networks, by using the network observed in the past. Our model outperforms the baseline model based only on pairwise interactions. Importantly, it also reveals how past interactions of a target hyperlink and of its neighbouring hyperlinks contribute to the prediction of the activation of the target hyperlink in the future.
This talk was developed in collaboration with Huijuan Wang (TU Delft) and Mathieu Jung-Muller (TU Delft, Orange).