A hybrid multi-group privacy-preserving approach for building decision trees

In this paper, we study the privacy-preserving decision tree building problem on vertically partitioned data. We made two contributions. First, we propose a novel hybrid approach, which takes advantage of the strength of the two existing approaches, randomization and the secure multi-party computation (SMC), to balance the accuracy and efficiency constraints. Compared to these two existing approaches, our proposed approach can achieve much better accuracy than randomization approach and much reduced computation cost than SMC approach. We also propose a multi-group scheme that makes it flexible for data miners to control the balance between data mining accuracy and privacy. We partition attributes into groups, and develop a scheme to conduct group-based randomization to achieve better data mining accuracy. We have implemented and evaluated the proposed schemes for the ID3 decision tree algorithm.