Credit Card Hash Representation for Searchable Fraud Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fraud detection systems face challenges in securely storing credit card numbers for analysis while maintaining the ability to identify patterns and trends, as encryption methods prevent the recognition of serial or issuer-related features.
Innovation Solution
Create a secure representation of a credit card number by hashing selected digits using a hash function, concatenating the prefix, position, and secret key, allowing for fraud detection without revealing the actual number.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If credit card numbers are encrypted to prevent fraudsters from obtaining the numbers, then security is improved, but fraud detection systems become unable to determine that two encrypted representations correspond to the same number or to identify serial relationships
Solution Approach 1:
The credit card number is divided into multiple digit groups, with different groups being hashed using different salt values. This segmentation allows the system to preserve enough information for fraud detection (by maintaining deterministic hashing within each group) while preventing reconstruction of the original number (by requiring multiple salted hash components).
Solution Approach 2:
A deterministic hashing function with salt values serves as an intermediary between the original credit card number and the stored representation. This intermediary transforms the number into a format that maintains pattern recognition capability (same input always produces same output) while preventing reverse engineering of the original number.
2Loss of information
If credit card numbers are stored in plain text to enable fraud detection analysis, then the ability to recognize patterns and trends is improved, but security risks increase as fraudsters can obtain the numbers if the database is breached
Solution Approach 1:
Instead of storing the original credit card numbers, the system stores deterministic hash representations as copies. These copies retain the essential property of identifiability (same number produces same hash) while eliminating the security vulnerability of storing actual sensitive data.
Solution Approach 2:
The system transforms the credit card number parameter into a hash value parameter through deterministic hashing. This parameter transformation maintains the functional requirement for pattern recognition (comparing hash values) while changing the data form to eliminate security risks associated with storing plain text numbers.
3Reliability
If existing encryption methods are applied to credit card numbers to protect security, then security is improved, but the encryption prevents fraud detection models from determining serial or issuer relationships between cards
Solution Approach 1:
By segmenting the credit card number into digit groups and applying different salt values to different groups, the system preserves structural information (prefixes indicating issuer or card type) in the hash output. This allows fraud detection models to identify serial relationships and issuer patterns while maintaining security through the salted hashing process.
Data Source
AI summary
Aspects of the present disclosure relate to creating a secure representation of a credit card number for fraud detection purposes. Embodiments include selecting a first set of digits of a credit card number, wherein the first set of digits comprises one or more digits of a prefix of the credit card number. Embodiments further include selecting a second set of digits of the credit card number. Embodiments further include creating a hash representation of each given digit of the second set of digits based on providing the given digit and the first set of digits as input to a hash function. Embodiments further include creating a secure representation of the credit card number based on the hash representations of each digit of the second set of digits.