Machine Learning Model Correction for Duplicate Item Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge lies in efficiently identifying and merging duplicate item description entries across different suppliers in network-accessible marketplaces, where variations in formatting and detail levels lead to significant time and resource consumption, and incorrect classification can result in false positives or negatives, affecting customer-merchant relationships.
Innovation Solution
A system and method utilizing machine learning techniques, such as genetic algorithms, support vector machines, and boosted decision trees, to generate rule sets for identifying and correcting marginal false positives by analyzing item description entries, distinguishing between duplicates and non-duplicates, and updating training data to improve model precision and recall.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If machine learning models are used to identify duplicate item description entries, then detection automation is improved, but false positive classification increases
Solution Approach 1:
The system implements feedback by identifying marginal false positives and using them to update training data for regenerating machine learning models. This closed-loop feedback mechanism continuously improves model accuracy by learning from its own errors, reducing false positives while maintaining automation.
Solution Approach 2:
The system performs preliminary action by identifying and correcting marginal false positives before they significantly degrade model performance. By proactively updating training data with corrected labels, the system prevents future misclassifications rather than merely reacting to them.
2Reliability
If traditional duplicate detection methods are used, then false positive classification is reduced, but time and computing resources are significantly consumed
Solution Approach 1:
The system applies partial action by focusing computational resources only on identifying and correcting marginal false positives rather than manually reviewing all duplicate detections. This selective approach maintains high reliability for critical cases while preserving overall productivity through automated processing of the majority of cases.
Solution Approach 2:
The system enables self-service by automatically identifying marginal false positives, updating training data, and regenerating models without requiring continuous human intervention. This self-correcting mechanism reduces the need for manual review while maintaining detection accuracy.
3Measurement precision
If machine learning models are regenerated frequently to improve precision, then detection accuracy is improved, but computational overhead increases
Solution Approach 1:
The system applies local quality by focusing model regeneration only on correcting marginal false positives rather than continuously regenerating models for all detection cases. This targeted approach improves detection accuracy for problematic cases while minimizing overall computational overhead by leaving well-performing model regions unchanged.
Data Source
AI summary
Embodiments of a system and method for identifying and correcting marginal false positives in machine learning models may include, based on reference data that includes pairs of information items and labels indicating whether pairs of information items have a specific relationship, generating a first machine learning model for determining whether pairs of information items have that relationship. Embodiments may include identifying one or more false positive pairs (e.g., a pair of information items that the first machine learning model indicates as having the specific relationship and which are labeled within the reference data as not having that relationship). Embodiments may include selecting identified false positive pairs as candidates for correction. Embodiments may include, subsequent to a correction of the reference data associated with the selected false positives, generating based on the corrected reference data a new machine learning model for determining whether pairs of information items have the specific relationship.