Project on anonymization of data, particularly related to temporal attacks
Given n anonymized tables deriving from the same private table, published at different points of time, the challenge is to ensure that no information can be obtained by studying these tables. The passage of time did not help the attacker gain any information about the entities.
Hypothesis
Given a private table PT, having N attributes totally, out of which Personal Identifiable attributes , such as name, address, phone number etc are removed, Q are quasi identifiers : these are attributes which when used in combination can be used to decode the identity of the individual, and S sensitive attributes, which could be disease in case of health records, income in case of census etc, there could be multi-sensitive attributes as well, but the main focus here is addressing a table with many quasi-identifiers and
DGH for DOB
D0 - dd/mm/yyyy
D1 - mm/yyyy
D2 - yyyy
D3 - in ranges of 10 → year
D4 - in ranges of 25
D5 - in ranges of 50
D6 - the century born, 2021, 2020 → [2000] and 1900-1999 → [1900]
17/03/2000
03/2000