Towards Generalization of Privacy Policy Specification and Property-Based Information Leakage

A decision-support framework for data anonymization with application to machine learning processes

“…The new RFD c i is then used to create a generalized dataset D 0 using the function COMPUTE_generalization, and the function COMPUTE_k is used to compute its anonymity level (lines [20][21]. If the anonymity level of D 0 is greater than the user-defined threshold t; c i is added to L and the data utility measures of D 0 are computed though the function COMPUTE_dataUtility (lines [22][23][24][25][26]. Finally, c i along with its anonymity level and data utility measures is added to R (line 27).…”

“…However, these applications have also introduced new privacy and ethical challenges [37]. Big Data typically holds large amounts of personally identifiable information (e.g., criminal records, shopping habits, credit and medical history, and driving records), which can enable mass surveillance and profiling programs and raise several privacy issues [23,18,8,9].…”