US 11,900,386 B2 · Methods and apparatus for payment transfer fraud detection

Abstract

This application relates to apparatus and methods for identifying fraudulent payment transfers. In some examples, a computing device determines payment transfer initiation features, and payment transfer reception features, based on previous payment transfer data. The computing device may train a machine learning fraud detection model with the payment transfer initiation features, and may train a machine learning fraud detection model with the payment transfer reception features. Once trained, the computing device may employ the machine learning fraud detection models to identify fraudulent payment transfers. For example, the computing device may determine whether a payment transfer is fraudulent when the payment transfer is initiated. Assuming the payment transfer is allowed, the computing device may determine whether the payment transfer is fraudulent when the payment is being received. In some examples, the computing device prevents completion of the purchase transaction if the payment transfer is determined to be fraudulent.

Claims20

Dependent claims

A system comprising:

a database;
a non-transitory memory;
a transceiver configured to receive a request for a payment transfer, wherein the request is stored in the non-transitory memory;
a processor configured communicatively coupled to the database and the non-transitory memory, wherein the processor is configured to:
obtain from the database, based on the received request obtained from the non-transitory memory, first data associated with previous payment transfers;
and
extract, from the request for the payment transfer and the first data, a first set of input features;
a fraud detection engine comprising at least one trained detection model configured to receive the first set of input features and generate second data including a fraud detection value based on the first set of input features;
a fraud determination engine configured to:
receive the second data;
determine whether the payment transfer is fraudulent based on a comparison of the second data to a fraud detection range;
based on the determination, generate response data identifying whether the payment is fraudulent;
store the response data in the database;
and
transmit, via the transceiver, in response to the request, the generated response data identifying whether the payment transfer is fraudulent.

12 more elements

A computer-implemented method, comprising:

receiving, via a transceiver, a request for a payment transfer;
obtaining, based on the request, first data associated with previous payment transfers;
and
extracting, from the request and the first data, a first set of input features;
applying, by a fraud detection engine, at least one trained detection model configured to receive the first set of input features and generate second data including a fraud detection value based on the first set of input features;
receiving, by a fraud determination engine, the second data;
determining, by the fraud determination engine, whether the payment transfer is fraudulent based on a comparison of the second data to a fraud detection range;
based on the determination, generating response data identifying whether the payment is fraudulent;
storing the response data in a database;
and
transmitting, via the transceiver, in response to the request, the generated response data identifying whether the payment transfer is fraudulent.

8 more elements

Dependencies are inferred from the claim text.

Methods and apparatus for payment transfer fraud detection

Life of this patent

Methods and apparatus for payment transfer fraud detection

Life of this patent

Abstract

Claims20