Performance Tuning of Steganography Algorithm for Privacy Preserving Association Rule Mining in Heterogeneous Data Base

Privacy and security issues in data mining become an important property in any data mining system. A considerable research has focused on developing new data mining algorithms that incorporate privacy constraints. In this paper, we focus on privately mining association rules in vertically partitioned data where the problem has been reduced to privately computing Boolean scalar products. We propose a modification of steganography-based multiparty protocols for this problem. The proposed modification fine tune the performance to be faster in case of very large database, with acceptable level of reduction in privacy.