ASENVA: Summarizing Anatomy Model by Aggregating Sensitive Values

Abstract

Anatomy model has been developed for reducing information loss on the previous model in the field of privacy-preserving data publishing. The previous model such as k-anonymity, p-sensitive, and l-diversity employ generalization and suppression to anonymize its data, while anatomy separate quasi identifier and sensitive attributes are separated into two tables. This separation potentially reduces information loss, but they ignore the records' efficiency. The number of records tends to increase due to the original table is divided into two. In this paper, a model is proposed to handle the efficiency problem of the number of records called ASENVA. The model aggregates sensitive values on a sensitive table and summarizes quasi identifier attributes. Therefore, the number of each record is reduced since each record on both tables is unique. A simple simulation with early simple generated data shows that our proposed model can reduce the number of records, indicating that our model is more efficient.