Identity Verification via Fuzzy Match Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing identity verification systems fail to accurately verify entities when there are inconsistencies in data, leading to multiple attempts and time-consuming searches across data sources.
Innovation Solution
An identity verification system generates a search structure with alternative queries based on subject attribute data, using wildcard tokens and character substitutions to accommodate variations, and determines a match score for candidate records to select the best match, even in the presence of errors or format differences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If exact match search is performed against data sources, then verification accuracy is improved, but verification fails when data inconsistencies exist leading to multiple attempts
Solution Approach 1:
The patent changes the search parameters from exact match to fuzzy match by introducing match score thresholds and similarity calculations. The system calculates match scores between candidate records and subject attribute data, allowing verification to succeed even when data inconsistencies exist, thereby resolving the contradiction between verification accuracy and verification success rate
Solution Approach 2:
The patent applies partial matching instead of requiring complete exact matches. By allowing partial matches above a certain threshold and ranking candidates by match score, the system can verify identities even with incomplete or slightly inconsistent data, improving verification success rate while maintaining reasonable accuracy
2Reliability
If multiple verification attempts are performed to handle inconsistent data, then verification completeness is improved, but time consumption increases
Solution Approach 1:
The patent performs preliminary fuzzy matching and ranking before final verification decisions. By pre-calculating match scores and ranking candidate records, the system identifies the most likely matches early, reducing the need for multiple verification attempts and thereby reducing time consumption while maintaining verification completeness
Solution Approach 2:
The patent replaces the mechanical process of multiple sequential verification attempts with an automated fuzzy matching and ranking system. The match score calculation and candidate ranking automatically handle inconsistent data without requiring manual retry cycles, significantly reducing verification time while maintaining completeness
3Adaptability or versatility
If fuzzy matching with wildcards is used to handle data variations, then search coverage is improved, but false positive matches increase
Solution Approach 1:
The patent uses match score feedback to filter and rank fuzzy match results. By calculating and comparing match scores against thresholds, the system can adjust the balance between search coverage and match accuracy, allowing broad search coverage while filtering out false positives through score-based ranking and selection
Data Source
Figure 1~5
Figure 2
Figure 3
AI summary
Embodiments of the present invention provide a computer-implemented method of verifying an identity of an entity against one or more data sources (131, 132, 133) wherein a record value is determined for each of a plurality of candidate records, determined based a score of each item of candidate attribute data associated with the record.