A Semantic-Based K-Anonymity Scheme for Health Record Linkage

Area of Science:

• Health Informatics
• Data Privacy
• Computer Science

Background:

• Record linkage is crucial for integrating sensitive data, especially medical records.
• Existing k-anonymity methods protect privacy but cause significant information loss through generalization.
• Individual-level data in record linkage remains vulnerable to re-identification despite anonymization.

Purpose of the Study:

• To propose a novel semantic-based linkage k-anonymity technique.
• To enhance de-identification in record linkage with reduced generalization.
• To eliminate inference disclosure risks through semantic reasoning.

Main Methods:

• Developed a semantic-based approach for k-anonymity in record linkage.
• Applied semantic reasoning to minimize data generalization.
• Evaluated the technique for privacy preservation and data utility.

Main Results:

• Achieved effective de-identification in record linkage.
• Demonstrated reduced information loss compared to traditional k-anonymity.
• Successfully eliminated privacy leakage through semantic inference.

Conclusions:

• Semantic-based linkage k-anonymity offers a superior method for de-identifying sensitive data.
• This approach balances privacy protection with data utility in record linkage.
• It provides a robust solution for secure medical data integration.