new method to find influential users based on interactions between pairs of users

Due to the increasing number of users in social networks, identifying a small number of users who can spread the information to a large group of people has become an important issue in social network analysis. According to the importance of this topic and its application in marketing, a lot of research has been done by scientists. In this study, a method is proposed called Node-Power-With-Neighbors (NPWN) to find influential users based on a feature for the node called Node-Power (NP) considering the power of the neighbors. We defined NP as a combination of normalized degree centrality and normalized shell-number of the node that is extracted from k-shell decomposition. We considered the amount of interactions between nodes as edge weights to determine the impact of neighbors on the nodes. To verify the effectiveness of our method, an epidemic information diffusion model is used called Susceptible-Infected-Recovered (SIR). We compared the proposed method with other basic methods on different datasets. The results show that NPWN outperforms other algorithms using Kendall’s tau coefficient