Duplicate Account Identification via Fuzzy Matching and Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing account management systems are ineffective in identifying and terminating duplicate accounts, as they rely on exact matches, fail to perform fuzzy matches, and cannot efficiently handle multiple accounts created for improper purposes, such as spam or penny stock scams.
Innovation Solution
A system that matches and scores user accounts based on various attributes, including user-supplied and system-detected information, generates a queue of potential duplicates, and prioritizes them for review, allowing for the termination of duplicate accounts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If exact match algorithms are used to identify duplicate accounts, then the system is simple to implement, but it fails to identify accounts with slight variations (fuzzy matches)
Solution Approach 1:
The system changes the matching parameters by implementing multiple matching algorithms with different stringency levels. The first algorithm uses strict exact matching on key fields, while the second algorithm uses fuzzy matching with configurable thresholds on various account attributes. This allows the system to adjust the precision of matching based on the specific context and account type.
Solution Approach 2:
The matching process is segmented into two distinct phases: first, an initial exact match phase that quickly identifies obvious duplicates, and second, a fuzzy match phase that handles more complex cases with variations. This segmentation allows the system to handle different matching scenarios with appropriate algorithms without requiring a single complex matching engine.
2Reliability
If the system reviews all matched account pairs, then identification completeness is high, but processing time increases significantly
Solution Approach 1:
The system applies partial action by using a two-stage approach where only account pairs that pass the first algorithm are subjected to the second algorithm. Additionally, the interface allows reviewers to focus on high-probability matches first, performing excessive action on critical cases while using automated termination for clear-cut duplicates, thus optimizing the balance between completeness and efficiency.
Solution Approach 2:
The system enables self-service through automated account termination capabilities. When the matching algorithms identify duplicate accounts with high confidence, the system can automatically terminate one of the accounts without requiring manual review, thus reducing processing time for obvious cases while maintaining identification completeness.
3Measurement precision
If multiple matching algorithms are implemented to improve identification accuracy, then false positives decrease, but system complexity increases
Solution Approach 1:
The system segments the matching functionality into distinct algorithms that can be independently configured and executed. Each algorithm focuses on specific matching criteria (exact match on key fields, fuzzy match on attributes, pattern recognition), allowing the complexity to be modularized and managed separately rather than as a monolithic complex system.
Solution Approach 2:
The system implements dynamics by allowing the matching algorithms to be configured and adjusted based on different account types and business requirements. The interface enables dynamic selection of which algorithms to apply and what thresholds to use, making the system adaptable to changing identification needs without requiring complete redesign.
4Productivity
If automated termination is implemented for clear-cut cases, then processing efficiency improves, but risk of false termination increases
Solution Approach 1:
The system implements feedback through a review interface where automated matching results are presented to users for verification. Users can review the matched account pairs, confirm or reject the automated decisions, and provide feedback that can be used to refine the matching algorithms. This feedback loop ensures that automated termination only occurs when the system is highly confident, reducing false terminations while maintaining efficiency.
Data Source
AI summary
A system matches accounts based on attributes of the accounts, and scores the matched account pairs based on a probability of the matched accounts being duplicate accounts. The system can utilize the matched and scored account pairs to determine duplicate accounts, and terminate at least one of the accounts in a duplicate account pair.