RumorDecay: Rumor Dissemination Interruption for Target Recipients in Social Networks

Abstract

Rumors (i.e., untrue emergence saying of COVID-19 in an area) that rapidly disseminate on the ubiquitous social media easily cause public panics and irrational behaviors (e.g., taking unnecessary medicine) of many very sensitive individuals referred to as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">target recipients</i> . Thus, rumor controlling or blocking for these target recipients is very critical, which differs from the traditional way of protecting all individuals and remains an open and challenging problem so far. In this work, on the basis that rumors are disseminated from the given sources to target recipients via multiple paths which may be significantly interrupted by deleting a few key links referred to as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">protectors</i> , we first mathematically define a general target information disseminating (TID) model and do theoretical proofs. Second, based on the TID model, we introduce a random walk algorithm to sample the paths of rumor dissemination for recipients. Third, aiming at deleting a budget-limited set of protectors efficiently in a large number of selected paths to reduce or weaken the negative rumor influences on the target recipients, we propose a heuristical strategy-based rumor influence decay mechanism referred to as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">RumorDecay</i> (i.e., RumorDecay <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> hop nearest neighbor method and RumorDecay <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> hop random walk method in this work) which can locate the optimal protectors quickly and efficiently. Finally, we conduct extensive experiments on many real social networks and the results show that the RumorDecay strategy can significantly weaken the rumor dissemination ability with less time cost.