Trust-based Misinformation Containment in Directed Online Social Networks

Publication Date


Document Type

Conference Proceeding

Publication Title

2023 15th International Conference on COMmunication Systems & NETworkS (COMSNETS)



First Page


Last Page



In today's world, Online Social Networks (OSNs) play a crucial role in our everyday life. But, its abuse to disseminate misinformation has turned out to be a major concern to us. Hence, the misinformation containment (MC) problem has attracted a lot of attention in recent times. For a given OSN with a fixed budget, this paper proposes a trust-based static technique independent of the distribution of misinformed nodes to select a set of trusted seed nodes leveraging the topologies of the network, to contain and decimate the misinformation faster. We follow a modified form of Competitive Linear Threshold Model with One Direction state Transition (LT1DT) to study the propagation dynamics of both the correct information and misinformation. Simulation studies on three real-world OSNs show that proposed method outperforms earlier work [1] significantly in terms of maximum number of misinformed nodes, infected time, point of inflection and number of misinformed nodes in steady state re-spectively. Moreover, its parallel implementation achieves almost 32 x speedup, making the procedure scalable for large scale OSNs to contain and decimate misinformation in real-time.


Competitive Linear Threshold Model, Disjoint and Overlapping Community Structures, General-Purpose Graphic-Processor Unit (GP-GPU), Misinformation Containment (MC), Online Social Networks (OSNs), Parallel Algorithms


Information Systems and Technology