False Positive Reduction Model for Fraud Detection Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current fraud detection methods in financial transaction monitoring often result in either false positives, leading to resource wastage and negative customer experience, or false negatives, allowing fraudulent transactions to occur.
Innovation Solution
A computer-implemented method using a server system that involves receiving payment transaction data and accuracy labels to train a series of False Positive Reduction Models (FPRMs) iteratively, allowing for the classification of transactions as high or low risk of being fraudulent, thereby reducing false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fraud detection models are trained to be aggressive in detecting fraudulent transactions, then fraud detection accuracy is improved, but false positive rate increases
Solution Approach 1:
The patent segments the fraud detection process into multiple specialized models: a primary fraud detection model for initial identification and a false positive reduction model for secondary verification. This segmentation allows each model to specialize in its specific function, improving overall accuracy while reducing false positives through divided responsibility.
Solution Approach 2:
The false positive reduction model acts as an intermediary between the primary fraud detection model and the final transaction decision. It receives transactions flagged by the primary model and performs additional analysis, serving as a mediator that reduces false positives before final classification.
2Reliability
If fraud detection models classify more transactions as high risk, then fraud prevention is improved, but resource wastage increases
Solution Approach 1:
The system segments transactions into different risk categories using specialized models, allowing resources to be focused on high-risk transactions that require intensive analysis while low-risk transactions are processed more efficiently.
Solution Approach 2:
The system applies partial action by using the false positive reduction model selectively on transactions that meet certain criteria, rather than applying exhaustive analysis to all transactions. This reduces resource wastage while maintaining fraud prevention effectiveness.
3Measurement precision
If fraud detection models are too aggressive, then fraud detection capability is improved, but customer experience deteriorates
Solution Approach 1:
The false positive reduction model serves as an intermediary that protects customer experience by catching and correcting aggressive classifications from the primary model before they reach the customer, thereby maintaining detection capability while improving experience.
Solution Approach 2:
The system provides beforehand cushioning by using the false positive reduction model to pre-filter and correct potential false positives before transactions are declined, cushioning the impact of aggressive detection on customer experience.
Data Source
AI summary
Systems and methods for reducing false positives for financial transaction fraud monitoring using machine learning techniques. Using an original model for separating transactions into high risk and low risk categories for fraud, transactions falling into the high-risk category may be labeled as false positive or true positive. The labels and data associated with the transactions may be used to train two or more sequential false positive reduction models (FPRMs). Once a desired minimal amount of false positive labels are found in transactions labeled high risk by one of the FPRMs, that one of the FPRMs becomes the trained FPRM. Next, an additional transaction is processed using the original model and the trained FPRM, which then determines whether the additional transaction is at a high risk of being fraudulent.