ARTIFICIAL NEURAL NETWORKS IN ACTION: SAFEGUARDING SOCIAL MEDIA FROM FAKE ACCOUNTS

Abstract

The rapid expansion of social media platforms has increased user engagement but also led to the proliferation of fake accounts, bots, and malicious profiles, undermining platform integrity and user trust. Detecting these fraudulent accounts is a complex challenge due to evolving deceptive behaviors and large-scale data streams. This study presents TrustNet, an artificial neural network (ANN)-based framework designed to identify and mitigate fake accounts on social media platforms. By analyzing user behavior patterns, profile attributes, and network interactions, the system trains a deep learning model to classify accounts as genuine or fake with high accuracy. The proposed framework incorporates feature selection, anomaly detection, and adaptive learning to handle dynamic social media environments. Experimental evaluations on benchmark social network datasets demonstrate that TrustNet outperforms traditional machine learning approaches in precision, recall, and overall detection accuracy, providing a scalable and reliable solution for maintaining platform security and trust.