Source Aware Budgeted Information Maximization

Abstract

The paper proposes a more general framework for budgeted influence maximization. We propose a novel cost function that considers the potential seed nodes' properties and the firm interested in maximizing the influence. A greedy algorithm, maximizing the influence to cost ratio, is then used to select a balanced set of seed nodes. We also show that the edge weights play an important role in determining the influential power of nodes. Further, the edge weights for a network can be efficiently predicted with the help of link prediction heuristics like resource allocation metrics and the Adamic-Adar index.