In this talk, I present recent work about algebraic measures of conflict in signed social networks. Signed networks are used to model conflict conflict in a social network, i.e., either direct negative relationships (represented by individual negative edges), or higher-level structures of conflict, such as a triangle consisting of two positive edges and one negative edge (I hate the friend of my friend.) While individual negative ties can be characterized as denoting conflict (in the sense that two persons are in conflict when they disagree), conflict can be defined in a broader way using balance theory: A signed graph is balanced when its nodes can be partitioned into two groups, such that all positive edges lie within each group and all negative edges connect the two groups. In this work, we introduce a measure of conflict ξ of a signed graph based on the signed graph's Laplacian matrix, by considering the relaxation of a frustration minimization problem.
29.10.2015 - 10:15