Distinct model on privacy protection of dynamic data publication

The M-Distinct is an excellent model that supports the anonymization of a fully dynamic set of data. This study aimed to explore and analyze the M-Distinct model. First of all, sensitive values in the QI-Group have certain randomness which caused the M-Distinct model prone to be property attacked. However, the (M, CUS)-Distinct model was proposed and required additional records in anonymity in the process. Therefore, its QI-Group-sensitive attribute value must belong to the same set of CUS to reduce the probability of property attacks. Secondly, the M-Distinct model involved time and cost. The proposed (M, CUS)-Distinct model creation phase generated disjoint barrel queues to ensure that each record could be stored in the record distribution phase, which reduced the time complexity of the algorithm. Finally, the experiment based on real data sets showed that the (M, CUS)-Distinct model was superior to the M-Distinct model in terms of data security, faked records processing, and execution time.