Auto-tuning Comparison Functions for Record Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing master data management systems face challenges in efficiently relating records from different systems with varying attribute structures and formats, leading to high computational costs and manual optimization efforts, which are time-consuming and resource-intensive.
Innovation Solution
A computer-implemented method and system that provides a set of comparison functions sorted by accuracy and computational cost, allowing for the selection of optimal functions based on predefined rules and resource constraints, enabling efficient record matching and false negative rate determination across different semantical classes, including encrypted records.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple comparison functions with high accuracy are used to relate records from different systems, then the matching precision is improved, but the computational cost increases significantly
Solution Approach 1:
The patent segments the comparison function selection process into distinct phases: first identifying attribute pairs between different record types, then selecting comparison functions for each attribute pair based on semantical classes. This segmentation allows the system to apply comparison functions selectively rather than uniformly across all attributes, reducing overall computational cost while maintaining matching precision for critical attributes.
Solution Approach 2:
The patent changes the parameter of comparison function selection by introducing accuracy thresholds and computational cost constraints. The system dynamically adjusts which comparison functions are applied based on these parameters, selecting higher-accuracy functions only when necessary to meet matching precision requirements while staying within computational budget limits.
2Measurement precision
If manual optimization of comparison functions is performed to achieve accurate record matching, then the matching accuracy is improved, but the time and resource investment increases
Solution Approach 1:
The patent implements self-service by enabling the system to automatically select and optimize comparison functions without manual intervention. The automated selection process uses predefined accuracy thresholds and computational cost constraints to determine the optimal set of comparison functions, eliminating the need for time-consuming manual optimization while maintaining high matching accuracy.
Solution Approach 2:
The patent performs preliminary action by pre-defining accuracy thresholds and computational cost constraints before the actual record matching process. Comparison functions are pre-selected and pre-configured based on these parameters, so that when record matching occurs, the system can immediately apply the optimal functions without requiring real-time manual optimization.
3Reliability
If a comprehensive set of comparison functions is applied to all attribute pairs, then the completeness of record matching is improved, but the device complexity increases
Solution Approach 1:
The patent applies local quality by selecting different comparison functions for different attribute pairs based on their specific characteristics and semantical classes. Instead of using a uniform comprehensive set of comparison functions across all attributes, the system tailors the selection to each attribute pair's requirements, maintaining matching completeness while reducing unnecessary complexity.
Solution Approach 2:
The patent segments the attribute space into different semantical classes and applies comparison functions selectively to each segment. This segmentation allows the system to maintain completeness of matching within each semantic category while avoiding the complexity of applying all possible comparison functions to all attributes uniformly.
Data Source
AI summary
A method for relating different types of records. The method may include providing comparison functions, wherein each comparison function corresponds to a semantical class, and wherein a computational cost is associated with each comparison function. The method may include determining one or more attribute pairs between the different types of records. The method may include sorting the comparison functions according to a determined accuracy. The method may include selecting a set of comparison functions associated with semantical classes according to a predefined rule. The method may include determining a total computational cost based on the computational cost of the selected set of comparison functions. The method may include determining whether two or more records are related using the selected set of comparison functions. The method may include relating the two or more records. The method may include determining a rate of false negative records.