To completely depersonalize the personal data of users on the Network, it is necessary to use a new technique, which was discovered by the developers of the Competence Center of the NTI “Technologies of Trusted Interaction”.
Ruslan Permyakov, Deputy Director of the Center at the Tomsk State University of Control Systems and Radioelectronics (TUSUR), told about this on March 29.
According to him, a special mathematical model of data quality assessment is used for depersonalization.
“It describes the boundary conditions for preserving the value of data for the developer and assesses the risk of re-identification”
As a result, when using the model, personal data does not become “too impersonal”, but it is still deprived of binding to a specific user. With the help of this model, developers select data coarsening parameters so that they can belong to several people.
“Usually, in the process of depersonalization, personal data becomes either too general, or it remains possible to link them to a specific person,” he explained.
The new method makes it impossible to correlate information with a specific user, the specialist stressed.
“Thanks to the development, we will be able to create a service that will form a dataset (database. — Ed.) according to the specified parameters,” he concluded.