Risk-Aware Data Anonymization via Pattern-Based Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data anonymization methods require significant human intervention and result in substantial information loss, as they generalize all records rather than focusing on records at risk of disclosure, failing to balance privacy with data utility effectively.
Innovation Solution
A risk-aware data anonymization method that identifies sensitive attributes and quasi-identifiers using association rule mining, computes record scores based on support and confidence metrics, and applies pattern-based K-Anonymization (PBKA) to cluster records and minimize information loss while maintaining privacy, using a greedy model to form m-sets and assess diversity metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing anonymization techniques generalize all records to a higher level, then privacy protection is improved, but information loss increases significantly
Solution Approach 1:
The patent applies local quality by differentiating treatment between high-risk and low-risk records. Instead of uniformly generalizing all records, the system identifies records with high re-identification risk and applies targeted anonymization only to those records, while preserving the original detail in low-risk records. This selective approach maintains privacy where needed while minimizing information loss overall.
Solution Approach 2:
The patent implements partial action by applying anonymization techniques only to the extent necessary for privacy protection. The system calculates risk scores for each record and applies generalization or suppression only to records exceeding a risk threshold, rather than applying uniform anonymization to all records. This partial application of anonymization reduces overall information loss while maintaining adequate privacy protection.
2Reliability
If existing anonymization methods are applied, then privacy protection is achieved to an extent, but human intervention is required significantly
Solution Approach 1:
The patent implements self-service by enabling the anonymization system to automatically identify high-risk records, calculate risk scores, and apply appropriate anonymization techniques without requiring manual review or intervention. The system autonomously performs association rule mining to identify quasi-identifiers, computes risk metrics, and executes the anonymization process, making the entire workflow self-operating while maintaining privacy protection.
Solution Approach 2:
The patent uses parameter changes by automatically adjusting anonymization parameters such as the level of generalization and suppression based on calculated risk scores. The system dynamically modifies these parameters for different records based on their risk profiles, enabling automated decision-making about the degree of anonymization needed without human intervention.
3Reliability
If existing anonymization techniques are used, then some privacy protection is provided, but data utility is compromised due to excessive generalization
Solution Approach 1:
The patent applies local quality by preserving data utility in low-risk records while ensuring privacy protection in high-risk records. The system maintains original record detail where re-identification risk is low, allowing meaningful analysis and computation on those records, while applying anonymization only where necessary to protect privacy.
4Reliability
If pattern-based K-Anonymization is applied to all records, then privacy is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent applies segmentation by dividing the record population into high-risk and low-risk segments based on risk score calculations. The system processes only the high-risk segment with computationally intensive pattern-based K-Anonymization techniques, while leaving the low-risk segment unchanged. This segmentation dramatically reduces computational complexity while maintaining privacy protection where it is most needed.
Solution Approach 2:
The patent implements partial action by applying the computationally demanding pattern-based K-Anonymization process only to records that exceed a risk threshold, rather than applying it universally to all records. This partial application reduces processing time and computational resources while maintaining adequate privacy protection for the most vulnerable records.
Data Source
AI summary
This disclosure relates generally to data anonymization and more particular y risk-aware data anonymization. Conventional data anonymization systems either replace PII/sensitive attributes with random values or shuffles them, that causes huge data distortion affecting the data utility. The goal of publishing data is best achieved when privacy is balanced with utility of data, Moreover, to ensure privacy, assessing the risk of disclosure is important. The proposed system provides a pipeline for analysis of data patterns to understand the associated risk level of re-identification of records. Further, based on the identified risks with the records the system anonymizes the data following a pattern based anonymization approach wherein data is clustered and for each cluster distinct patterns are identified such that the information loss is minimal.