Bidirectional Entity Matching for Higher-Quality Recommendations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Interpreting large volumes of customer and service provider data to provide meaningful recommendations is complicated and time-consuming, particularly when determining suitable matches between entities.
Innovation Solution
A computer-implemented method that determines a matching score for candidate and target entities by considering bidirectional suitability measures, using a system configured to maintain accounting records and infer attributes from network data, allowing for improved matching suggestions based on historical records and actual work experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data collection and interpretation methods are used to provide meaningful recommendations, then recommendation quality is improved, but complexity and time consumption increase
Solution Approach 1:
The matching process is divided into distinct segments: extracting attributes from entity data, computing suitability measures separately for target and candidate entities, and combining these measures to generate matching scores. This segmentation simplifies the overall complexity by breaking down the interpretation process into manageable, modular steps that can be executed independently.
Solution Approach 2:
The system transforms raw entity data into standardized parameters (attributes) that can be systematically compared. By changing the representation of data from unstructured text to structured parameters with defined weights, the system improves recommendation quality while maintaining computational efficiency through consistent parameter-based evaluation.
2Measurement precision
If data collection and interpretation methods are used to provide meaningful recommendations, then recommendation quality is improved, but time consumption increases
Solution Approach 1:
The system performs preliminary actions by pre-extracting and storing entity attributes in a standardized format, and by pre-defining suitability measure calculations. When matching is needed, these pre-prepared components are quickly combined and evaluated, significantly reducing the time required for real-time recommendation generation while maintaining high recommendation quality.
3Measurement precision
If bidirectional suitability measures are calculated for entity matching, then matching accuracy is improved, but computational complexity increases
Solution Approach 1:
The system employs asymmetric evaluation by calculating suitability measures in two different directions: target suitability (how suitable the target entity is for the candidate) and candidate suitability (how suitable the candidate entity is for the target). This asymmetric approach improves matching accuracy by considering both perspectives, while the modular calculation method keeps computational complexity manageable.
Solution Approach 2:
The bidirectional suitability measures act as counterweights to each other, balancing the evaluation process. By computing both target suitability and candidate suitability and combining them through a defined function, the system achieves more accurate matching results while the structured combination method prevents exponential growth in computational complexity.
4Measurement precision
If multiple entity attributes are considered for matching, then matching quality is improved, but processing complexity increases
Solution Approach 1:
The system transforms multiple diverse entity attributes into a unified parameter framework where each attribute is represented as a weighted suitability measure. This parameter transformation allows the system to consider multiple attributes simultaneously while maintaining processing efficiency through consistent mathematical operations on standardized parameters rather than handling raw attribute data.
Solution Approach 2:
The system discards redundant or less important attributes by applying weighting factors that reduce their influence on the final matching score. Less critical attributes are effectively discarded in the calculation, while more important attributes are recovered and given appropriate weight, thereby improving matching quality without proportionally increasing processing complexity.
Data Source
AI summary
A method comprises identifying a candidate entity for matching with one or more target entities from a plurality of target entities, and for one or more of the target entities: determining a target suitability measure indicative of a suitability of the target entity to the candidate entity based on desired target entity attributes of the candidate entity and corresponding one or more target entity attributes associated with the target entity; determining a candidate suitability measure indicative of a suitability of the candidate entity to the target entity based on one or more desired candidate entity attributes of the target entity and corresponding one or more candidate entity attributes associated with the candidate entity; determining a matching score for the candidate entity and target entity pair as a first function of the target suitability measure and the candidate suitability measure; and. determining suggested target entities for matching with the candidate entity.