Influence Maximization Problem With Echo Chamber Effect in Social Network

An echo chamber effect describes the situation in which opinions are amplified by communication and repetition inside a relatively closed social system. In this article, we will detect the echo chamber effect in real-world data set and measure this effect during the information diffusion process. Any user will be influenced by its neighbors or echo chamber effect. Also, we assume that these activation events from each activated neighbor and from echo chamber effect are independent. In this article, we detect and model the echo chamber effect for the first time. Then, the influence maximization with echo chamber (IMEC) problem aims to select $k$ users to propagate information such that the expected number of activated users is maximized. We formulate this problem using a graph model and analyze the NP-hardness. Second, the objective of IMEC as a set function is proved to be neither submodular nor supermodular. Then, an improved greedy algorithm is proposed, which is combined metaheuristic strategies. Finally, experimental results show that our algorithm is effective in detecting echo chamber effect and efficiency in selecting seed nodes.