Models of Social Influence: A Survey
Résumé
Online Social Networks (OSN) have become an integral part of human life in the world
today. Social media applications have permeated almost every aspect of our lives. The concept
of social influence has emerged from social networks and has brought with it stimulating
research interest areas like influence maximization, viral marketing and sentiment analysis.
Most researchers provide specic contexts for analyzing influence such as topics, opinions or
intensity of interactions. As a result many different views exist of what constitutes influence
on social network. In general, empirical social influence research has three phases: collection
of social network data, formation of a social graph out of the collected data and use of the
social graph to compute influence scores through specific metrics. While many surveys on
Social Influence exist, there is hardly a survey that has taken an approach that explores
various techniques and metrics that have been used particularly to abstract the social graph
and to calculate influence out of the social graph. In this survey, we review past and current
approaches that have been applied in each of these phases of social influence study over time.
We also review some attributes of social network data and preprocessing activities undertaken
to make such data suitable for social influence analysis. The review concludes with a summary
of current trends in social influence research, existing open issues for further investigation and
future opportunities for exploitation.
Origine : Fichiers produits par l'(les) auteur(s)