Enhancing data privacy: Predictive analytics for detecting malicious user activity in social networks
Sailaja Terumalasetti, S R Reeja · 2024
The prediction of malicious users from the perspective of social networks is an indispensable and significant theme in the context of data confidentiality of individuals. Cryptography, ML, DL, and AI are prevailing elucidations for destructive user identification and prediction. The exploration commences with an innovative method to enhance data privacy in predictive analytics, aiming to identify fraudulent user behaviour in OSN. The paper presents an innovative framework that combines DRL with Fuzzy Logic to effectively tackle the issues of achieving accurate predictions while preserving data privacy. The methodology anticipates attaining robust identification of malicious behaviours by integrating DRL algorithms for learning effective detection strategies with fuzzy logic for handling uncertainty in OSN data. It ensures the confidentiality of critical user information. The conclusions outperform in protecting the privacy of data.