Privacy Preserving Association Rule Mining on Distributed Healthcare Data: COVID-19 and Breast Cancer Case Study.

Abstract

Association rule mining can be used in healthcare data mining to provide solutions to life-threatening diseases like recent COVID-19. Due to healthcare data privacy concerns, privacy preserving distributed healthcare data mining becomes the primary focus of medical science research. Recently, Chahar et al. (Sādhanā 42:1997–2007, 2017) proposed privacy preserving distributed association rule mining scheme with insecure communication channels. They used the concept of an elliptic curve-based paillier cryptosystem to achieve privacy, authenticity, and integrity. We observed some security vulnerabilities in their privacy preserving association rule mining scheme when implemented with insecure communication channels. We observed that the security vulnerabilities will result in the disclosure of private data of sites (or participants). Furthermore, we propose a secure version of their scheme to solve the security vulnerabilities with insecure communication channels. Theoretical and experimental analysis shows that the proposed scheme has almost equal computation and communication complexities with better securities. A case study on the effectiveness of the proposed approach in combating COVID-19 coronavirus and Breast Cancer is also discussed.