VGMFSN: Design of an Efficient Fused VARMA GRU Model for the Identification of Fake Profiles on Social Networks

Social networks, which provide a platform for communication and information sharing, have become integral to people’s daily existence. However, as the use of social networks has increased, the prevalence of false profiles has become a serious concern. These fraudulent profiles can injure individuals, businesses, and society. Consequently, identifying fake personas on social networks has become an important endeavor. In this study, we propose an effective VARMA-GRU fusion model for detecting false profiles on social networks. Combining the VARMA (Vector Autoregressive Moving Average) model and the GRU (Gated Recurrent Unit) model improves the classification task’s accuracy. The VARMA model captures the intricate temporal dependencies between the features of the false profiles.In contrast, the GRU model represents the sequential behavior of the profiles. We compile a fraudulent and authentic social network profile dataset to evaluate the proposed model. Regarding accuracy, precision, recall, and F1 score, the experimental results demonstrate that the proposed model outperforms current methodologies. The accuracy of the proposed model is 98.5%, which is substantially higher than the accuracy of other methodologies. The proposed fused VARMA GRU model is a highly efficient and precise method for identifying false profiles on social networks. The model can assist social network platforms in enhancing their security and safeguarding their users from malicious activities.