Privacy Preserving Association Rule Mining Revisited

Abstract: The privacy preserving data mining (PPDM) has been one of the most interesting, yet challenging, research issues. In the PPDM, we seek to outsource our data for data mining tasks to a third party while maintaining its privacy. In this paper, we revise one of the recent PPDM schemes (i.e., FS) which is designed for privacy preserving association rule mining (PP-ARM). Our analysis shows some limitations of the FS scheme in term of its storage requirements guaranteeing a reasonable privacy standard and the high computation as well. On the other hand, we introduce a robust definition of privacy that considers the average case privacy and motivates the study of a weakness in the structure of FS (i.e., fake transactions filtering). In order to overcome this limit, we introduce a hybrid scheme that considers both privacy and resources guidelines. Experimental results show the efficiency of our proposed scheme over the previously introduced one and opens directions for further development.