Grouping of Distributed Data at Multiple Sites by Sustaining the Privacy of Data

Data with respect to individuals are available at multiple sites. Essential inferences can be made based on the collective information that is distributed. However, privacy of the data maintained be different organizations becomes a major concern. There is a necessity to construct precise models to mine crucial information from the distributed data without placing the data in a centralized storage. Existing work concentrates on manipulating the data at each of the sites and then clusters them. But it is observed that maneuvering the huge amount of data is expensive. The proposed work discusses a privacy preserving clustering method to group identical data. One of the major challenges while modeling is to compute a representative for each cluster. A secure approach is also proposed to calculate the cluster representative. When compared to the existing privacy preserving distributed clustering approaches, the proposed method provides better accuracy and reduced error rate.