Table 7 Final re-identification risk results after the application of privacy-preserving techniques to the quasi-identifiers attributes.
From: Empowering open data sharing for social good: a privacy-aware approach
Transformation | Generalisation (range 5) | Noise ([−3, 3]) | Generalisation (centrality - quartiles) | — | Generalisation (re-coding category) | Re-identification risk (%) | Metric |
|---|---|---|---|---|---|---|---|
QI attributes | Age | DateOfFirstPositiveLabResult | HospitalisationDays | Gender | Outcome | ||
✓ | ✓ | 0 | k-anonymity | ||||
✓ | ✓ | ✓ | 0.89 | k-anonymity | |||
✓ | ✓ | ✓ | 1.07 | k-anonymity | |||
✓ | ✓ | ✓ | ✓ | 3.38 | k-anonymity | ||
✓ | ✓ | ✓ | 13 | Record linkage | |||
✓ | ✓ | ✓ | ✓ | 0 | Record linkage | ||
✓ | ✓ | ✓ | ✓ | ✓ | 0 | Record linkage |
