Effect of Influential Nodes on Time Varying Opinion Formation Models

Eeti, Anurag Singh
2020 Procedia Computer Science  
Information diffusion is considered as ubiquitous phenomena. There can be various points of information flow in a social network e.g, social contacts, media. People are connected through their social contacts to construct a social network. Social network is never static as connections keep on changing among the people. Continuously changing network is termed as Temporal or Time varying network. Probability of constructing new connections is higher in comparison to removal of connections.
more » ... connections. Information flows among the network creates opinion among the people. Opinions of people keep on effecting the others opinion which are neighbours of that person . This is how opinion dynamics work. There are cases of fake news, or spreading good will related to a product or campaigning in politics. Influential nodes play a major role in all these applications. Some high influential nodes according to their location in the network are picked and bribed to spread some information among the network creating a specific required opinion among the social network. In this work, time varying opinion formation models are considered, considering different properties e.g, opinion difference threshold, weighted opinions to consider different types of models e.g, simple model, weighted opinion model, weighted opinion threshold model. It is analyzed that adding which property, influential nodes can work better and influence the network quickly. Opinion Difference Threshold is considered to check the difference of opinions between two neighbouring nodes. Nodes are weighted according to their influence to calculate the weighted opinion. On these models, effect of influential nodes is analyzed. It is also discussed that how much percentage of influential nodes can bring the network to consensus to the required opinion trend. Abstract Information diffusion is considered as ubiquitous phenomena. There can be various points of information flow in a social network e.g, social contacts, media. People are connected through their social contacts to construct a social network. Social network is never static as connections keep on changing among the people. Continuously changing network is termed as Temporal or Time varying network. Probability of constructing new connections is higher in comparison to removal of connections. Information flows among the network creates opinion among the people. Opinions of people keep on effecting the others opinion which are neighbours of that person . This is how opinion dynamics work. There are cases of fake news, or spreading good will related to a product or campaigning in politics. Influential nodes play a major role in all these applications. Some high influential nodes according to their location in the network are picked and bribed to spread some information among the network creating a specific required opinion among the social network. In this work, time varying opinion formation models are considered, considering different properties e.g, opinion difference threshold, weighted opinions to consider different types of models e.g, simple model, weighted opinion model, weighted opinion threshold model. It is analyzed that adding which property, influential nodes can work better and influence the network quickly. Opinion Difference Threshold is considered to check the difference of opinions between two neighbouring nodes. Nodes are weighted according to their influence to calculate the weighted opinion. On these models, effect of influential nodes is analyzed. It is also discussed that how much percentage of influential nodes can bring the network to consensus to the required opinion trend.
doi:10.1016/j.procs.2020.06.016 fatcat:j2cgxna3gvapvc7fkgqlfa57le