Adaptive Transaction Processing Using Machine Learning Peer Groups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current transaction processing systems face limitations in dynamically identifying potentially anomalous activity, often resulting in false positives and failing to detect elusive transactions due to reliance on static criteria and manual processing, which is laborious and prone to human error.
Innovation Solution
The system employs adaptive transaction processing techniques using machine learning to classify entities into peer groups based on shared attributes, applying trained transaction models to evaluate transaction data contextually and compute prioritization indicators, thereby reducing false positives and improving detection of anomalous activity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If static criteria and manual processing are used to evaluate transaction data, then the system is simple to implement, but the accuracy of identifying potentially anomalous activity deteriorates due to false positives and inability to detect elusive transactions
Solution Approach 1:
The patent implements dynamic transaction processing by continuously learning from transaction data to adapt evaluation criteria. The system transitions from static rules to dynamic models that evolve with new data, improving detection accuracy while managing complexity through automated learning processes.
Solution Approach 2:
The system changes parameters by using machine learning models that automatically adjust evaluation thresholds and criteria based on learned patterns. This allows the system to adapt sensitivity and specificity parameters dynamically, reducing false positives while maintaining detection capability.
2Adaptability or versatility
If static criteria are used to evaluate all transactions, then the processing rules are easy to implement, but the adaptability to different entity types and contexts deteriorates
Solution Approach 1:
The patent segments transactions into peer groups based on entity attributes and transaction characteristics. This segmentation allows different evaluation criteria to be applied to different segments, improving adaptability while managing complexity through localized rule sets rather than universal complex rules.
Solution Approach 2:
The system dynamically adapts evaluation criteria based on the specific peer group and context of each transaction. Rather than using fixed static rules, the system adjusts its evaluation approach based on learned patterns specific to different entity types, enhancing versatility.
3Productivity
If manual processing is used to review transactions, then the system requires minimal automation, but the productivity and throughput deteriorate due to laborious processing and human error
Solution Approach 1:
The system performs self-service by automatically learning from transaction data and improving its own evaluation capabilities without continuous manual intervention. The machine learning models autonomously adapt to new patterns, reducing the need for manual rule updates while maintaining high productivity and accuracy.
Solution Approach 2:
The system implements feedback loops where transaction outcomes and reviews are used to continuously improve the machine learning models. This feedback mechanism enables automated processing to become increasingly accurate over time, maintaining high throughput while reducing errors that would require manual correction.
Data Source
AI summary
Systems and techniques are described for applying machine learning techniques to dynamically identify potentially anomalous activity of entities. In some implementations, peer group data is obtained. The peer group data indicates multiple entities classified as belonging to a particular peer group, and a set of attributes associated with the multiple entities. Transaction data for the multiple entities is obtained from one or more data sources. One or more transaction models are selected. The transaction models that are each trained to apply a particular set of evidence factors corresponding to the set of attributes associated with the multiple entities, and identify transaction patterns representing potentially anomalous activity. The transaction data is processed using the one or more transaction models to identify potentially anomalous activity within the transaction data for the multiple entities. A prioritization indicator is computed for each entity included in the multiple entities.