Μη γραμμικά μοντέλα διαμόρφωσης γνώμης στα κοινωνικά δίκτυα
Non-linear opinion-formation models on social networks
View/ Open
Keywords
Μη γραμμικά μοντέλα ; Non linear models ; Κοινωνικά δίκτυαAbstract
Social networks have a significant impact on people's behavior and opinion. For this reason, in
order to investigate the role of these networks in social behavior, many mathematical models of
social networks have been developed in the past. Research on the dynamics of public opinion
mainly follows two types of models, the models based on statistical physics and the models based
on data. The first type of model is traditionally designed to capture many real-life regulatory
phenomena. However, these models are almost computerized and therefore their parameters are
difficult to learn from detailed real data. The second category of models, which is surprisingly few,
attempts to overcome these limitations by learning a linear detection model from the transient
dynamics of thought. Most of these approaches do not consider predicting an arbitrary future time
to evaluate the utility of their models. Instead, they focus on near prediction in order to predict the
next time stamp opinion. This paper studies the opinion-forming models, with emphasis on nonlinear
models, for which a bibliographic review is performed, while also presenting the analysis of
the Non-linear DeGroot algorithm where a simulation of a system with two agents was performed.