SANATOMY: Privacy Preserving Publishing of Data Streams via Anatomy

Compared with generalization, anatomy preserves both the privacy and the correlation in data publication. On the other hand, data streams have gradually become a widely used data representation. Therefore, in this paper, we develop a novel algorithm of SANATOMY, to solve the problem of anatomized publishing of data streams. It creates l-diverse buckets according to the stream tuples´ sensitive values, and controls the maximum release delay of each tuple. It also merges part of the buckets or re-partitions all the tuples into new buckets, while the bucket cannot be published straight. Experiments show that our algorithm allows significantly more effective data analysis than generalization in data streams, and has a better performance on data real-time processing and utilization.