Anonymization Apparatus Reducing Information Loss via Recursive Data Division
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The k-anonymity method for anonymizing personal data often results in significant information loss, leading to erroneous determinations when analyzing anonymized data, as excessive data conversion is required to ensure privacy protection.
Innovation Solution
An anonymization apparatus that determines the appropriate attributes for division and recursively divides personal data into groups, ensuring that attribute values within each group become identical, thereby reducing information loss through a Mondrian method and Loss Metric evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If k-anonymity method is applied to anonymize personal data, then privacy protection is improved, but information loss increases
Solution Approach 1:
The patent segments the anonymization process into multiple stages: first dividing the data into groups based on quasi-identifier attributes, then applying k-anonymity within each group. This segmentation allows selective anonymization that preserves more information while still achieving privacy protection.
Solution Approach 2:
The patent applies different anonymization strategies to different groups of data. By first clustering data into groups with similar characteristics and then applying k-anonymity locally to each group, the method preserves local patterns and reduces overall information loss compared to global anonymization.
2Object-affected harmful factors
If data conversion operations are repeated to achieve k-anonymity, then privacy protection is improved, but information accuracy deteriorates
Solution Approach 1:
The patent performs preliminary grouping of data into clusters before applying k-anonymity. This preliminary action organizes the data structure in advance, allowing subsequent anonymization operations to work within pre-defined groups, thereby reducing the number of repeated conversions needed and preserving information accuracy.
Solution Approach 2:
The patent employs feedback mechanisms to evaluate the quality of anonymized data and adjust the anonymization process accordingly. By measuring information loss and privacy protection levels, the system can optimize the anonymization depth to achieve the minimum necessary conversions for privacy while maintaining information accuracy.
Data Source
AI summary
According to one embodiment, an anonymization apparatus according to an embodiment is configured to execute a determination process, a division process, and a process of recursively executing at least the determination process and the division process, and to thereafter execute anonymization. A number-of-kinds calculation circuit calculates a number of kinds of different attribute values for each of a plurality of attributes, before the determination process is executed. A determination circuit determines the attribute to be noticed, based on also the calculated number of kinds. A sort circuit sorts a plurality of personal data items, based on the calculated number of kinds, before the division process is executed.