An efficient and effective hop-based approach for influence maximization in social networks

摘要: Influence maximization in social networks is a classic and extensively studied problem that targets at selecting set of initial seed nodes to spread the influence as widely possible. However, it remains an open challenge design fast accurate algorithms find solutions large-scale networks. Prior Monte Carlo simulation-based methods are slow not scalable, while other heuristic do have any theoretical guarantee they been shown produce poor for quite some cases. In this paper, we propose hop-based can be easily applied billion-scale under commonly used Independent Cascade Linear Threshold diffusion models. Moreover, provide provable data-dependent approximation guarantees our proposed approaches. Experimental evaluations with real network datasets demonstrate efficiency effectiveness algorithms.