Predicting signs and directions of links in online social networks

Abstract

Studying online social networks gives us a new way to visualize the network data. The relations between the users in the social networks can be de ned as positive and negative where positive means friendship and negative relations means antag- onism. Signs in the social networks are important because the attitude of one user toward another user can be estimated from evidence provided by their common relationships and also other members surrounded in the speci c social network. We studied how the relationships can be denoted by the signs. For that purpose we need to predict the sign (relationship type) between two users where the users can be represented by the nodes and their relationship can be represented by signed edges. Again implementing triad and a new theme quad has been implemented for better demonstration to predict the signs and directions between actors of social networks.