Towards a Two-Tier Architecture for Privacy-Enabled Recommender Systems (PeRS)

The current surge in recommender systems research is impressive, but it has highlighted a number of concerns, including users’ privacy and data security. Although various solutions to these privacy breaches have been proposed, the existing solutions fall short of directly addressing the real issues, and most of them continue to rely on third parties. Moreover, giving third parties access to users’ personally identifiable information (PII) is a cause for concern. In this paper, we have suggested a two-tiered architecture. The identity of the users is anonymized in the first tier, in the next tier, homomorphic encryption is exploited for the purpose of randomization of the users’ data. With the help of proposed solution, on one hand, third-party involvement can be eliminated. On the other hand, there can be improvement in the privacy mechanism by providing a two-tier architecture for user data security. The suggested framework is intended to serve as a baseline for safeguarding users’ privacy and integrity when conducting online purchases and other associated activities.