Granular Transaction Data Obfuscation via Noise Injection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Granular transaction data is valuable but often cannot be shared due to privacy concerns, as it reveals sensitive information about individuals and entities involved in transactions.

Innovation Solution

Introducing noise into transaction data to obfuscate it, while ensuring that the modified data remains useful and accurate for analysis, by adjusting or swapping transaction values and adding/removing data entries, thus maintaining privacy while retaining the original metric values within a predetermined range.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If granular transaction data is shared at detailed level, then data usefulness and analytical value are improved, but privacy of individuals and entities is compromised

Engineering Contradiction:

Improvedata usefulnessVSAvoidprivacy compromise

Core Design Contradiction:

Loss of informationVSObject-affected harmful factors

Solution Approach 1:

A processing server acts as an intermediary between the transaction data source and requesting entities. The server receives data requests, applies noise injection to obfuscate granular details, and returns modified data that preserves analytical value while protecting privacy. This intermediary process enables data sharing without direct exposure of sensitive information.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transforms transaction data by changing its parameters through noise injection. Specifically, it adjusts transaction amounts, adds or removes transactions, and modifies timestamps by random amounts within defined ranges. These parameter changes obfuscate individual transaction details while maintaining aggregate statistical properties useful for analysis.

Inventive Principle:

Principle #35Parameter changes

2Object-affected harmful factors

If noise is injected to obfuscate transaction data, then privacy is protected, but data accuracy may deteriorate

Engineering Contradiction:

Improveprivacy protectionVSAvoiddata accuracy

Core Design Contradiction:

Object-affected harmful factorsVSMeasurement precision

Solution Approach 1:

The patent applies partial noise injection rather than complete obfuscation. It injects noise into only certain parameters (transaction amount, timestamp, merchant information) while preserving other critical data elements. The noise magnitude is controlled to remain within acceptable ranges that maintain data utility for analytical purposes.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system incorporates feedback mechanisms to monitor the quality of obfuscated data. It evaluates whether the injected noise maintains data within acceptable ranges for analytical use, allowing adjustment of noise parameters to balance privacy protection with data accuracy requirements.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS10181050B2Method and system for obfuscation of granular data while retaining data privacy

Publication Date: 2019.01.15 MASTERCARD INT INC

AI summary

A method for obfuscating granular transaction data via the introduction of noise includes: storing transaction data entries, each including transaction data values including at least a transaction amount and merchant identifier; receiving a data request including selection criteria and desired metrics; identifying a subset of transaction data entries based on the selection criteria; identifying a metric value for each desired metric based on the corresponding transaction data value in each of the transaction data entries in the subset; inserting noise to modify the subset by (i) adjusting the transaction data values included in at least two of the transaction data entries, and/or (ii) adding at least one transaction data entry to and removing at least one transaction data entry from the subset, where the inserted noise adjusts the metric values within a predetermined range; and transmitting the modified subset.