Time-sensitive propagation values discount centrality measure

Abstract

The detection of influential individuals in social networks is called influence maximization which has many applications in advertising and marketing. Several factors including propagation delay affect the degree to which an individual influences the network. Many different methods, including centrality measures, identify high-influence individuals in social networks. The time-sensitive harmonic method (TSHarmonic), which considers time sensitivity to propagation delay and duration, is a centrality measure. TSHarmonic has two weaknesses: high computational complexity and ignoring the influence of the selected node in selecting other influential nodes. Therefore, in this article, the valuable path-finding process in the TSHarmonic method is modified to provide the Fast Time-Sensitive Harmonic algorithm. The provided method has the same accuracy as the TSHarmonic, while the speed is significantly increased. Then, the Time-Sensitive Propagation Values Discount method is proposed to improve detection speed and accuracy. This method takes into account the influence of the selected node for future selection and hence increases the accuracy.