Entity Resolution System with Justification Engine
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current record matching technologies lack justification for match outcomes, leading to inefficiencies and inaccuracies in the verification process, which can result in incorrect record linkages and compromised data security.
Innovation Solution
An entity resolution system that generates and stores deterministic rules based on attribute patterns, providing human-readable justifications for match outcomes, enabling verifiers to focus on specific data fields and improving the speed and accuracy of verification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If probabilistic matching technologies are used to resolve entities, then record linkage can be performed across disparate provider organizations, but accuracy of record linkage deteriorates due to errors in demographic data collection and transcription
Solution Approach 1:
The patent introduces an intermediary verification process that acts as a mediator between probabilistic matching results and final record linkage. Human verifiers review and validate the probabilistic matches, correcting errors caused by demographic data issues. This intermediary step preserves interoperability while improving accuracy by filtering out false matches caused by data collection and transcription errors.
2Measurement precision
If verification process is implemented to validate probabilistic matching outcomes, then accuracy of record linkage is improved, but verification efficiency deteriorates due to lack of match justification data
Solution Approach 1:
The patent implements feedback by providing verifiers with detailed match justification data that explains why records were matched or not matched. This feedback includes information about matching fields, similarity scores, and reasoning behind the probabilistic matching outcomes. Verifiers can then quickly validate or reject matches without manually reviewing all data elements, significantly improving verification efficiency while maintaining high accuracy.
3Device complexity
If binary match results are provided without justification, then system complexity is reduced, but verification accuracy deteriorates due to human error in reviewing all data elements
Solution Approach 1:
The patent segments the verification process by providing verifiers with focused information about specific matching fields and their significance rather than requiring review of all data elements. The match justification data breaks down the matching rationale into discrete, reviewable components such as individual field comparisons and their weights. This segmentation reduces the cognitive load on verifiers and minimizes human error while maintaining system simplicity.
Data Source
AI summary
Generating and providing match justifications in association with record matching results for improved record matching is provided. An entity resolution system generates a set of deterministic rules based on patterns of attribute comparison outcomes of known matched and unmatched records. Each rule includes matching conditions and an action instruction including a message of a justification for a match/non-match. The system receives a request to determine a match between two records. A matching engine compares various attributes of the records using probabilistic matching technologies to determine a match outcome. A justification engine compares attribute comparison outcomes to the stored rules. When the matching conditions of a rule are satisfied, the rule is activated and a human-readable justification is linked to the match output and provided in a match response. The justification provides insights into the match outcome, and improves the speed and accuracy of a verification process of the match output.