Data Privacy Preserving Recommendations for Social Media

To deliver personalized services to users, recommendations systems are highly important. Despite the numerous advantages, implementing personalized recommendation systems usually necessitates the collection of users' personal data, especially online usage, which unnecessarily exposes users to significant privacy concerns. As a result, developing a functional privacy-preserving framework to prevent sensitive information from being inferred against inference attacks is critical. The main idea is to mask the users' information such that the original data is transformed with data anonymization to comply with data security and privacy regulation. As a result, using different entropy-based metrics, our system will quantify privacy leakage, which provides effective privacy protection for individuals.