Back to results
US 11,238,528 B2Granted
Systems and methods for custom ranking objectives for machine learning models applicable to fraud and credit risk assessments
AssigneeAmerican Express Travel Related Services Company, Inc.
InventorVishwa Vinay
Life of this patent
Estimating the term…
Abstract
A system for analyzing risk using machine learning models may be trained using a data set to generate a risk assessment model that is optimized for metrics commonly used in for financial risk evaluation. The metrics may include Gini and CaptureRate, for example. The system may receive a request for a financial service, and generate a risk assessment by applying the risk assessment model to factors associated with the request. The system may also decide on the request in response to the risk assessment.
Claims17
A method comprising:
- training, by a computing device, a modeling module to detect and prevent fraudulent online activity associated with at least one merchant, said training comprising performing a continuous, iterative validation of electronic transaction data via an executed loss function algorithm based at least on a Gini metric derived from an area under curve (AUC) plot and a capture rate metric of labeled data within the electronic transaction data, wherein the training further comprises:
- swapping, by the computing device, a first value and a second value from one metric of the Gini metric and the capture rate metric in response to the first value being lower than a predetermined threshold;
- dynamically generating and updating, by the computing device, via the trained modeling module, a risk assessment model, such that the risk assessment model is updated in accordance with the training of the modeling module;
- receiving, by the computing device, a request for an electronic transaction via an electronic financial service;
- generating, by the computing device, a risk assessment by applying the generated risk assessment model to factors associated with the electronic transaction;
- and
- deciding, by the computing device, on the request in response to the risk assessment, said decision triggering a point-of-sale (POS) device associated with the electronic financial service to accept or deny the request.
4 more elements
A computer-based system, comprising:
- a processor;
- and
- a tangible, non-transitory memory configured to communicate with the processor, the tangible, non-transitory memory having computer-executable instructions stored thereon that, in response to execution by the processor, cause a computerized risk analysis system to perform operations comprising:
- training the computerized risk analysis system to detect and prevent fraudulent online activity associated with at least one merchant, said training comprising performing a continuous, iterative validation of electronic transaction data via an executed loss function algorithm based at least on a Gini metric derived from an area under curve (AUC) plot and a capture rate metric of labeled data within the electronic transaction data, wherein the training further comprises:
- swapping, by the computerized risk analysis system, a first value and a second value from one metric of the Gini metric and the capture rate metric in response to the first value being lower than a predetermined threshold;
- dynamically generating and updating, via the computerized risk analysis system, a risk assessment model, such that the risk assessment model is updated in accordance with the training of the computerized risk analysis system;
- receiving, by the computerized risk analysis system, a request for an electronic transaction via an electronic financial service;
- generating, by the computerized risk analysis system, a risk assessment by applying the risk assessment model to factors associated with the electronic transaction;
- and
- deciding, by the computerized risk analysis system, on the request in response to the generated risk assessment, said decision triggering a point-of-sale (POS) device associated with the electronic financial service to accept or deny the request.
7 more elements
An article of manufacture including a non-transitory, tangible computer readable storage medium having instructions stored thereon that, in response to execution by a risk analysis system, cause a computing device to perform operations comprising:
- training the risk analysis system to detect and prevent fraudulent online activity associated with at least one merchant, said training comprising performing a continuous, iterative validation of electronic transaction data via an executed loss function algorithm based at least on a Gini metric derived from an area under curve (AUC) plot and a capture rate metric of labeled data within the electronic transaction data, wherein the training further comprises:
- swapping, by the computing device, a first value and a second value from one metric of the Gini metric and the capture rate metric in response to the first value being lower than a predetermined threshold;
- dynamically generating and updating, via the risk analysis system, a risk assessment model, such that the risk assessment model is updated in accordance with the training of the risk analysis system;
- receiving, by the risk analysis system, a request for an electronic transaction via an electronic financial service;
- generating, by the risk analysis system, a risk assessment by applying the risk assessment model to factors associated with the electronic transaction;
- and
- deciding, by the risk analysis system, on the request in response to the generated risk assessment, said decision triggering a point-of-sale (POS) device associated with the electronic financial service to accept or deny the request.
4 more elements
Dependencies are inferred from the claim text.