UNIVERSAL PROCESS MODEL FOR SECURE DATA DE-IDENTIFICATION WITH DYNAMIC VERIFICATION

The process of data depersonalization, which is a key mechanism for reducing regulatory requirements and minimizing privacy threats, is often implemented as a disparate set of technical procedures without a unified control loop. This leads to inconsistencies between the methods used, the actual level of residual risk, and the target indicators of data utility for subsequent processing. This article proposes a universal process model for de-identification, the Unified De-identification Process Model (UDPM), which formalizes this process as a closed control loop with dynamic assessment of the risk-utility trade-off. At the core of the model is a feedback mechanism that performs iterative adjustment of the parameters of the methods used, based on a quantitative assessment of the risk of re-identification and utility metrics