Constrained information flows in temporal networks reveal intermittent communities

Area of Science:

• Network Science
• Complex Systems
• Data Science

Background:

• Real-world networks often represent dynamic systems with interactions changing over time.
• These dynamic systems can be modeled as multilayer networks, with each layer representing a temporal snapshot.
• Accurate links between layers are crucial but often unknown, leading to simplistic assumptions in current methods.

Purpose of the Study:

• To develop a principled approach for estimating node-level layer dependencies in multilayer temporal networks.
• To improve the identification of intermittent community structures and model information spreading.

Main Methods:

• A novel method for estimating node-level layer dependencies based on intra-layer network structure.
• Implementation within the Infomap community detection framework.
• Evaluation on synthetic and real temporal networks.

Main Results:

• The proposed node-level coupling method effectively constrains information within multilayer communities.
• Improved recovery of planted groups in benchmark networks.
• Enhanced identification of intermittent communities in real temporal contact networks.

Conclusions:

• Node-level layer coupling is superior to current methods for analyzing temporal networks.
• This approach enhances the modeling of information spreading and community detection in dynamic systems.
• The method offers a more accurate representation of real-world temporal network dependencies.