Differentially private data release based on clustering anonymization

Abstract

Based on the theory of anonymization,the DBSCAN method was applied to divide all the data records into different groups to cover individuals.To provide priv enhancement,the Laplace noise was added to the anonymized partitioned data to perturb the real value of data record so that the requirements of differential privacy model were satis-fied.With the clustering operation,the sensitivity of the query function has been partitioned to improve data utility.The proof of privacy has been given and experimental results have been provided to evaluate the utility of the released data.

Keywords

• differential privacy;privacy preservation;clustering;data release;anonymization