For Dynamic Data Set Re-release Of Privacy Protection

As the emergence and development of data applications such as data publishing and data mining,a critical challenge is to preserve data privacy and prevent sensitive information disclosure.Recently,the anonymization technique has become the highlight of privacy preserving research community.Because it can preserve privacy while guarantee the truth of the published data,these characteristics make it applicable for the applications from various areas.Most existing anonymization work has been done on static dataset,which has no update and need only one-time publication.Recent studies consider anonymizing dynamic dataset with external updates:the dataset is updated with record insertions and/or deletions.However, anonymizing dynamic dataset is more challengeable in contrast to the anonymization of static dataset.Firstly,the dynamic dataset will be updated continually as time evolves,thus it is usually required to be anonymized and re-published in different time.Secondly,in plenty of practical applications,the dynamic dataset will be updated by internal updates:the attribute values of each record in the dataset are dynamically updated.This paper investigates the anonymization work of fully dynamic dataset,which has both of external updates and internal updates.Using the typical existing anonymization techniques, we show the invalidation of existing methods.We introduce formal definitions and theoretical analysis of dynamic dataset,and present a general privacy disclosure framework that is applicable to all anonymous re-publication problems.We propose a new counterfeited generalization principle called m-Distinct to effectively anonymize dataset with both external updates and internal updates.We also develop an algorithm to generalize datasets to meet m-Distinct.The experiments conducted on real-world data demonstrate the effectiveness of the proposed solution.