Online Social Network‐Mediated Rumor Propagation Among Cyber Consumers and Its Consequences on E‐Commerce: A Mathematical Model

ABSTRACT The swift spread of online social network rumors creates huge challenges for e‐commerce stands, as purchaser conviction and the market can be permanently disrupted. This study proposed the ICRSbGCN, which is established on the ICRS (Ignorant, Confused, Rumor‐monger, and Stifler) and the use of Graph Convolutional Networks (GCNs) to examine rumor propagation among cyber consumers. By adding confused consumers, the ICRSbGCN model accounted for more complexity or nuance to consumer behavior while also emulating the consumer reservation behaviors seen in the real‐world context in considering misinformation. To derive the epidemic thresholds, equilibrium points, and stability conditions, the principles behind rumor spread dynamics were utilized. The results from numerical simulations demonstrate that rumors disappear when the basic reproduction number (Ro) is less than 1, while an infectious threshold with Ro greater than 1 reflects as endemic in cyberspace. These results also demonstrated the utility of the ICRSbGCN model to identify the critical influencers in an e‐commerce domain and reduce misinformation. The ICRSbGCN model achieved 99.10% accuracy, 99.23% precision, 99.15% recall, and 99.43% F1‐score while outperforming previously known techniques. The importance of targeting misinformation is emphasized specifically to protect consumer trust and uphold the e‐commerce sphere through the proposed ICRSbGCN model. Overall, the ICRSbGCN model shows an accurate method of analyzing and reducing rumor spread in online social networks through finding influencers and predicting the misinformation spread across many states of a consumer.