A temporal network version of Watts's cascade model

TL;DR

This paper extends Watts's cascade model to temporal networks, incorporating a finite influence window, and tests it on empirical data to better understand social contagion dynamics.

Contribution

It introduces a temporal aspect to Watts's cascade model by including a finite influence window and compares two threshold types using real datasets.

Findings

Temporal influence window affects cascade spread.

Model performs well on empirical datasets.

Threshold type impacts cascade dynamics.

Abstract

Threshold models of cascades in the social sciences and economics explain the spread of opinion and innovation due to social influence. In threshold cascade models, fads or innovations spread between agents as determined by their interactions with other agents and their personal threshold of resistance. Typically, these models do not account for structure in the timing of interaction between the units. In this work, we extend a model of social cascades by Duncan Watts to temporal interaction networks. In our model, we assume friends and acquaintances influence agents for a certain time into the future. That is the influence of the past ages and becomes unimportant. Thus, our modified cascade model has an effective time window of influence. We explore two types of thresholds -- thresholds to fractions of the neighbors or absolute numbers. We try our model on six empirical datasets and compare them with null models.