Practical anonymous subscription system with privacy preserving data mining

Abstract

To date, one interesting research topic in constructing anonymous subscription systems is how to allow client profiling, while keeping customers anonymous when they access one service. Though several solutions have been proposed, the service providers are only endowed with limited ability of utilizing and analyzing accumulated transaction transcripts at the cost of weakened privacy protection. To overcome this obstacle, we put forth the first anonymous subscription system with privacy preserving data mining, which is derived by applying the technique of Kiayias-Xu-Yung data mining group signature to the underlying multi-service subscription system by Canard and Jambert. The most prominent benefit of the new system is that service providers can obtain the desired output by a quorum of trusted data mining servers, and at the same time the customers can preserve maximum possible anonymity. Performance comparison shows that the proposed system is more practical than several related schemes published recently.