Time Blocking Noising for Data De-identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current de-identification methods, such as those under HIPAA standards, are insufficient to prevent re-identification of individuals in de-identified data sets, as patterns of activity in traditionally de-identified data can still be matched with external sources, allowing for subject re-identification.
Innovation Solution
The method involves identifying bursts of temporally-proximate events in data sets, which are then randomly shifted to preserve temporal relationships while obfuscating identifying features, making the data unmatchable with external sources, thereby preventing re-identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If traditional de-identification methods (such as HIPAA standards) are used to remove identifying features from data sets, then data privacy is improved, but the data can still be re-identified by matching patterns of activity with external sources
Solution Approach 1:
The patent segments the data set into multiple time blocks based on temporal proximity of events. Each time block is independently shifted by a random time amount, breaking the continuous temporal pattern that could be used for re-identification while preserving the internal temporal relationships within each block. This segmentation approach prevents matching with external sources while maintaining data utility.
Solution Approach 2:
The patent changes the temporal parameter of event timestamps by adding random time shifts to each time block. This parameter transformation obfuscates the absolute timing information that could be used for re-identification, while preserving the relative temporal relationships within each block. The random time shift parameter effectively breaks the link between de-identified and external data sources.
2Ease of manufacture
If all date data is removed to meet HIPAA de-identification standards, then compliance is improved, but the temporal relationships needed for analytics are lost
Solution Approach 1:
The patent applies different treatments to different parts of the data: within each time block, temporal relationships are preserved to maintain analytics utility, while between time blocks, random shifts are applied to remove identifying patterns. This local differentiation allows the data to be both compliant with de-identification standards and useful for temporal analytics.
Solution Approach 2:
The patent performs preliminary segmentation of the data into time blocks before applying de-identification transformations. This preliminary organization allows subsequent random shifts to be applied in a controlled manner that preserves internal temporal relationships while breaking external matchability, achieving both compliance and analytics utility.
3Reliability
If random time shifts are applied to time blocks, then re-identification is prevented, but the absolute timing information is altered
Solution Approach 1:
The patent introduces dynamic random time shifts to each time block, making the de-identified data adaptable and resistant to static matching attacks. The random shifts create a dynamic transformation that prevents re-identification while preserving the relative temporal structure within blocks, balancing privacy protection with analytics utility.
Data Source
AI summary
Techniques disclosed herein relate to removing potentially identifying features of a specific subject from a data set to prevent re-identification of the subject using an external data source. In various embodiments, the data set contains, as potential identifying features of the specific subject, multiple bursts of temporally-proximate events. Time blocks within the data set can be identified to capture one or more of the bursts of temporally-proximate events for the specific subject. Adding random time shifts for each time block can add noise to the data set and remove or obfuscate the identifying features of a specific subject to generate a time shifted data set.