Fairness Budget Distribution for Influence Maximization in Online Social Networks

Abstract

Influence maximization ( $${\mathsf {IM}}$$ ) is an important problem in social influence, viral marketing, and economics. This paper studies a fairness constraint in the Influence Maximization problem, a general $$ {\mathsf {IM}}$$ version that aims to find a k-size seed set distributed in target communities. Each has certain upper and lower bounds so that the influence spread is maximal. However, solving this problem faces two main challenges: it is an NP-hard problem, and the seed sets with strong influence may not satisfy the fairness constraint. To address this problem, we propose the Fairness Budget Influence Maximization algorithm. This algorithm combines an improved greedy strategy with generating sampling and a stop-and-stare technique. We conducted experiments on real social networks. The result shows that our proposed algorithm equalizes or outperforms the state-of-the-art algorithm in terms of the objective value, the running time, the memory usage, and especially the target communities coverage ratio of the seed set. It is almost greater than the compared algorithm and guarantees fairness constraint.