Distributed Model Based Sampling Technique for Privacy Preserving Clustering

The sharing of data has been proven beneficial in data mining applications. However, privacy regulations and other privacy concerns may prevent data owners from sharing information for data analysis. To resolve this challenging problem, data owners must design a solution that meets privacy requirements and guarantees valid data clustering results. To achieve this dual goal, we introduce a new method for privacy-preserving clustering based on the probability distributed model for clustering. This paper proposes square wave-based clustering model, gauss distribution-based clustering model and the multivariate normal distribution-based clustering model based on the information provided by the K-means clustering result separately. And we mainly show the correctness and feasibility of the sampled data for clustering by experiment. This new preserving technique can be used not only for distributed clustering, but also for simultaneous clustering rule and data hiding.