Source-linked AI summary
Dynamical and bursty interactions in social networks
Juliette Stehle, Alain Barrat, Ginestra Bianconi
TL;DR
Empirical social-contact data show broad contact durations and inter-contact intervals, motivating models that represent burstiness and memory. This paper models rapidly evolving group contacts and finds that its rules can produce broad or narrow distributions, with stationary behavior limited to specified parameter conditions.
Problem
Social-contact data show broad contact durations and inter-contact intervals, while modeling work in this area was still developing simple frameworks that reproduce these patterns.
Source: p. 1Method
The paper introduces a simple, analytically and numerically tractable model of rapidly evolving, disconnected social groups, with agents joining or leaving according to state- and time-dependent rules.
Source: pp. 1–2Results
The model produces narrow or broad contact-duration distributions; with reinforcement dynamics, broad distributions emerge, and simulated state-time distributions agree with analytical predictions and remain broad under changed microscopic rules.
Source: pp. 2–3Takeaways & Limitations
The framework can be extended to model social-contact mechanisms and used to simulate processes such as information spreading on rapidly evolving networks.
Source: p. 4The stationary analytical results require b1 > 1/2, λ > 1/2, and b0 > (2λ − 1)/(3λ − 1); outside these boundaries, stationarity is violated.
Source: p. 3
Abstract
from arXiv · showhide
We present a modeling framework for dynamical and bursty contact networks made of agents in social interaction. We consider agents' behavior at short time scales, in which the contact network is formed by disconnected cliques of different sizes. At each time a random agent can make a transition from being isolated to being part of a group, or vice-versa. Different distributions of contact times and inter-contact times between individuals are obtained by considering transition probabilities with memory effects, i.e. the transition probabilities for each agent depend both on its state (isolated or interacting) and on the time elapsed since the last change of state. The model lends itself to analytical and numerical investigations. The modeling framework can be easily extended, and paves the way for systematic investigations of dynamical processes occurring on rapidly evolving dynamical networks, such as the propagation of an information, or spreading of diseases.