Quantifying the effect of temporal resolution on time-varying networks

Area of Science:

• Network Science
• Complex Systems
• Dynamical Processes

Background:

• Time-varying networks model systems with evolving interactions.
• These networks are often simplified into static snapshots over time intervals (Δt).
• This simplification can distort the analysis of dynamic processes.

Purpose of the Study:

• To quantify the impact of time aggregation (Δt) on dynamical processes in time-varying networks.
• To develop a mathematical framework for understanding this impact.
• To improve the characterization of dynamics on evolving graphs.

Main Methods:

• Focusing on the random walk as an elementary dynamical process.
• Developing a mathematical framework to describe the bias introduced by time aggregation.
• Validating the framework with real-world datasets.

Main Results:

• Time aggregation with arbitrary time intervals (Δt) demonstrably biases dynamical processes.
• A mathematical framework accurately describes the observed behavior on real datasets.
• The study provides analytical insights into the introduced bias.

Conclusions:

• The choice of time aggregation interval (Δt) significantly affects the analysis of dynamics on time-varying networks.
• The developed framework enables a more accurate characterization of these dynamics.
• This work advances the understanding of temporal network analysis.