Simple Distribution of Sensitive Values for Multiple Sensitive Attributes in Privacy Preserving Data Publishing to Achieve Anatomy
Widodo Widodo, Murien Nugraheni, Irma Permata Sari · 2021
In anonymizing data, k-anonymity is a pioneer model and become popular. However, it still a drawback in information loss when the sensitive values are not distributed evenly. This study aims to distribute evenly sensitive values in microdata with multiple sensitive attributes by extending simple distribution of sensitive values (SDSV) method to anatomy. Previously, this method works well when it is conducted in k-anonymity. This method is used with a little modification in last step by exchange randomly record if privacy level is not satisfied and this method is run in anatomy. The result shows that in anatomy SDSV has better performance than systematic clustering as a base line. SDSV has diversity value 1.02 and systematic clustering has 0.33. From information loss aspect in anatomy model the result also shows that SDSV outperforms systematic clustering.