Predicting the evolution of social networks with life cycle events

Abstract

This paper presents a model of social network evolution, to predict and simulate changes in social networks induced by lifecycle events. We argue that social networks change with lifecycle events, and we extend a model of friendship selection to incorporate these dynamics of personal social networks. The model uses theories of homophily and reciprocity and is formulated in a random utility maximization framework to predict the formation of social ties between individuals in the population. It is then extended to predict the evolution of social networks in response to life cycle events. The model is estimated using attribute data of a national sample and an event-based retrospective dataset collected in 2009 and 2011 respectively. Findings suggest that homophily has a strong effect on the formation of new ties. However, heterophily also plays a role in maintaining existing ties. Although the motivation of this research stems from incorporating social network dynamics in large-scale travel behaviour micro-simulation models, the research can be used in a variety of fields for similar purposes.

Determinants of Social Networks in Rural China: Does Transportation Have a Role to Play?

“…The importance of mobility is realized only when individuals are connected and meet with each other (Urry, ). Sharmeen, Arentze, and Timmermans () showed that proximity to highways is not associated with denser social networks, whereas proximity to train stations increases personal networks. More puzzling, their study suggested that the number of nearby shopping facilities positively influence an individual's social interactions.…”

Social interactions in transportation: analyzing groups and spatial networks

“…Two of these articles are motivated by a goal of incorporating social network dynamics, and decisions made in inter-and intra-household networks, in travel behavior simulation models. Sharmeen et al (2015) develop a dynamic life cycle event model of friendship using a random utility maximization framework to predict the formation of a tie between two individuals. Using a Dutch dataset, the authors find that homophily as well as geographic proximity play an important role in the formation of friendship networks.…”

“…Dubernet and Axhausen (2015) draw on co-evolutionary principles from natural computing to develop an algorithm to efficiently search approximate solutions for their framework for joint decision making, which in turn is inspired by game theory. Sharmeen et al (2015) propose a model of social network evolution inspired by life course theory and concepts from social network analysis, in a novel cross-fertilization from sociology, which is then brought to a utility-maximizing framework. In a comprehensive multi-agent based social simulation model, Okushima (2015) uses results from combined revealed and stated preference estimation, a hierarchical Bayesian model and two hazard-based duration models.…”

Modelling the loss and retention of contacts in social networks: The role of dyad-level heterogeneity and tie strength

“…A key characteristic of social networks is that they are not static but evolve over time. Some first attempts to model social network dynamics and their interaction with life-cycle events have been made by (Chávez et al, 2017;Sharmeen et al, 2014Sharmeen et al, , 2015Sharmeen et al, , 2016. While these papers have modelled the changes in social networks over time, more can be done to accommodate the extent of the heterogeneity involved in this process at the respondent and at the dyad level.…”