Locating multiple rumor sources in social networks using partial information of monitors

Rumors in social media platforms and the identification of their sources is a challenging issue in modern-day computer communication. Existing approaches mostly fail to localize the source node accurately due to the lack of complete network information or timestamps. Besides, most of the techniques focused on single-source identification only, while sometimes multiple sources exist in the network. In this paper, we designed a new algorithm called Multi Snowballing with Partial Timestamps (MSPT) to find multiple sources utilizing partial timestamps available to monitors. We have explored the snowballing technique to determine the vulnerable radius that may contain the rumor source based on the partial timestamps of a few nodes. The overall complexity of the algorithm is $O(N_S\ast (N_S+E_S))$, where $N_S$ is the set of snowball nodes and $E_S$ represents edges in between snowball nodes. Extensive empirical analysis is performed on a variety of networks, which include small-scale, large-scale, and artificial networks. Empirical outcomes demonstrate that the presented algorithm is efficient in terms of error distance and execution time compared to baseline algorithms.