Fraud Detection Ensemble Reduces Dominant Feature Bias
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fraud detection models in transactions are overly sensitive to dominant features, leading to inaccurate fraud scores due to oversensitivity.
Innovation Solution
A system and method that balance features by combining the results of multiple fraud detection models, each trained on different feature sets, to generate a more accurate fraud score.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single fraud detection model uses all available features including dominant features, then the model can capture comprehensive fraud indicators, but the model becomes overly sensitive to dominant features leading to inaccurate fraud scores
Solution Approach 1:
The patent divides the fraud detection task into multiple separate models, each trained on a different subset of features. This segmentation prevents any single dominant feature from overwhelming the entire detection system, as each model only considers a portion of the total feature set. The ensemble combination of these segmented models achieves comprehensive fraud detection while maintaining score reliability.
Solution Approach 2:
The patent extracts and removes dominant features from specific model training sets to prevent oversensitivity. By deliberately excluding these influential features from certain models and relying on other models to contribute to the overall detection, the system prevents any single feature from dominating the fraud score calculation.
2Measurement precision
If multiple fraud detection models are used to balance features, then the accuracy of fraud detection improves, but the system complexity increases
Solution Approach 1:
The patent combines multiple fraud detection models into an ensemble system where each model processes a subset of features. The results from individual models are merged through a scoring mechanism that weights and aggregates their outputs. This merging approach maintains improved accuracy while managing complexity through structured integration.
Solution Approach 2:
The ensemble of fraud detection models serves multiple functions simultaneously: each individual model detects specific fraud patterns based on its feature subset, while the collective ensemble provides comprehensive fraud detection. This multi-functionality allows the system to handle diverse fraud scenarios without requiring separate specialized systems for each case.
Data Source
AI summary
A system and method for improving the accuracy of fraud detection in transactions for a more effective fraud response. First and second models each receive transaction request data, evaluate the transaction request data for fraud, and produce first and second initial fraud results, respectively. The first model includes a first set of features which is relevant to detecting fraud and which includes one or more dominant features. The second model includes a second set of features which excludes the one or more dominant features. A score generator receives and combines the first and second initial results to generate a final fraud score. A response module takes an action based on the final score, such as rejecting or allowing the transaction request. The first and second models may use the same or different machine learning or other modeling technologies. The score generator may be a combination module or a combination model.