A Novel Framework for Privacy Enabled Healthcare Recommender Systems

The proliferation in the research of recommender systems especially in the healthcare domain raised up several problems particularly security and privacy of patients’ medical data. Majority of the existing work depends on third parties that manage healthcare systems. Access of the PII, i.e., “personally identifiable information,” of the patients to the health-managing third parties is a sincere issue. Although various suggestions to these privacy violations have been made, the present solutions lack in addressing the primary issues. In our paper, we have proposed a two-tier framework to discuss the same. Firstly, the patient’s identity is anonymized, and then a homomorphic encryption mechanism is used to randomize data of patients. The main purpose of the given framework is twofold. On one hand, it is independent of the third-party involvement; on the other side it also improves privacy by consolidating two-layer architecture for the patients’ data security. The present work adds the privacy mechanism in the recommendation research and also projected that the recommended solution can help as a standard for the protection of patients’ privacy and integrity.