Fake Account Detection Using Machine and Deep Techniques in Social Media

Detecting fake accounts is vital in safeguarding Online Social Network (OSN) platforms from malicious activities, ensuring genuine user interactions, and preventing issues like misinformation and fraud. The pervasive nature of fake accounts poses challenges such as social manipulation, disruption of democratic processes, and erosion of online discourse credibility. Businesses also face consequences, including skewed analytics and financial losses. This study employs advanced machine learning algorithms Ada-boost, Cat-boost, Extra'Tree, Long Short Term Memory network (LSTM), Bidirectional Encoder Representations from Transformers (BERT), and Autoregressive Mode (XLnet) to enhance fake account detection accuracy. The performance of the proposed methods is evaluated using different performance metrics such as accuracy, precision, recall and Fl-score and finally the results are compared. The results shows that Cat-boost is outperforming in all these metrics compared to all other methods.