privacy-record-linkage
Installation
SKILL.md
Privacy-Preserving Record Linkage
Overview
Privacy-Preserving Record Linkage (PPRL) enables two or more organizations to identify matching records across their datasets without revealing the underlying personal data to each other. This is critical for healthcare research, fraud detection, national statistics, and cross-organizational analytics where direct data sharing is prohibited by privacy regulations.
Approach Comparison
| Approach | Privacy Level | Accuracy | Scalability | Communication Cost |
|---|---|---|---|---|
| Bloom Filter Encoding | High | Good (>95% F1) | Very High | Low |
| Secure Hash Matching | Very High | High (exact match only) | Very High | Very Low |
| Secure Multi-Party Computation | Cryptographic | Very High | Medium | High |
| Trusted Third Party | Depends on TTP | Very High | High | Medium |
| Differential Privacy Linkage | Formally private | Moderate | High | Low |
Bloom Filter-Based PPRL
How It Works
Related skills