The K-Anonymization Method Satisfying Personalized Privacy Preservation

Even if k-anonymity model can prevent publishing data from disclosing privacy effectively and efficiently, due to the uneven distribution of the sensitive data, ordinary k-anonymization method cannot guarantee each tuple satisfying the personalized privacy requirement of it's data owner although the publishing table has been satisfied k-anonymity constraint. The reason which k-anonymity table fails to satisfy personalized privacy requirement is analyzed firstly, then Correlate degree of Sensitive Values, Leakage Collection, privacy disclosure metric and data quality metric are presented. At last an anonymization method satisfying personalized privacy requirements is presented, in which a utility-driven adaptive clustering method is proposed to partition tuples with similar best data quality.