# US 2025/0156869 A1 · Systems and methods for efficient progressive fraud detection

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0156869 A1
- Document ID: US20250156869A1
- Status: Published application
- Kind: A1
- Application number: 18/505,767
- Filed: 2023-11-09
- Published: 2025-05-15
- Earliest claimed: 2023-11-09
- Assignees: Stripe, Inc.
- Applicants: Stripe, Inc.
- Inventors: Ryan Lee Drapeau; Dulwin Eksith Jayalath
- Web page: https://patentgrep.com/patent/US20250156869A1

## CPC

- G06Q 20/4016: {involving fraud or risk level assessment in transaction processing}
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 5/025: Extracting rules from data
- G06N 5/046: Forward inferencing; Production systems
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/00: Machine learning
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A method and apparatus for efficient and progressive fraud detection are described. Transactions are received by a server computer system and include attributes and transaction data. To determine whether a transaction is fraudulent, and thus should be rejected, a progressive fraud determination process is performed by the server computer system. The progressive fraud determination process includes performing a first set of one or more transaction fraud determinations based on a first set of features determined for the transaction. When the first set of one or more transaction fraud determinations identifies the transaction as fraudulent or legitimate, the progressive fraud determination process is terminated based on this decision, and before performing additional successive fraud determination processes. By exiting the progressive fraud determination process early, substantial processing, memory and bandwidth savings can be realized without sacrificing fraud detection accuracy.

## Claims

### Claim 1

A method for fraud detection performed by a server computer system, the method comprising:
receiving, by the server computer system, a transaction from a user having transaction attributes and transaction data;
initiating, by the server computer system, a progressive fraud determination process that determines whether the transaction is fraudulent or legitimate;
performing, by the server computer system, a first set of one or more transaction fraud determinations based on a first set of features determined for the transaction;
in response to a determination that the first set of one or more transaction fraud determinations has identified the transaction as fraudulent or legitimate, terminating, by the server computer system, the progressive fraud determination process;
in response to a determination that the first set of one or more transaction fraud determinations is inconclusive of determining whether the transaction is fraudulent or legitimate, performing, by the server computer system, a second set of one or more transaction fraud determinations based on a second set of features determined for the transaction, wherein the second set of features is greater than the first set of features; and
terminating, by the server computer system, the progressive fraud determination process based on a fraud determination rendered by the second set of one or more transaction fraud determinations that identifies the transaction as fraudulent or legitimate.

### Claim 2 (depends on 1)

The method of claim 1, wherein each of the second set of one or more transaction fraud determinations determines whether the transaction is legitimate, the method further comprising:
in response to a determination that the second set of one or more transaction fraud determinations identify the transaction as legitimate, performing, by the server computer system, a third set of one or more transaction fraud determinations based on a third set of features determined for the transaction, wherein the third set of one or more transaction fraud determinations determines whether the transaction is fraudulent, and wherein the third set of features is a maximum set of transaction features that is greater than the second set of features; and
terminating, by the server computer system, the progressive fraud determination process to process or reject the transaction based on a fraud determination rendered by the third set of one or more transaction fraud determinations that identifies the transaction as fraudulent or legitimate.

### Claim 3 (depends on 2)

The method of claim 2, wherein terminating the progressive fraud determination process after the first set of one or more transaction fraud determinations, the second set of one or more transaction fraud determinations, or the third set of one or more transaction fraud determinations comprises:
transmitting an instruction to a transaction management system to process the transaction or reject the transaction based on a fraud determination result generated by the first set of one or more transaction fraud determinations or the second set of one or more transaction fraud determinations.

### Claim 4 (depends on 2)

The method of claim 2, wherein the first set of one or more transaction fraud determinations comprises a set of static rule fraud determinations, wherein the second set of one or more fraud determinations comprises at least one machine learning model trained to identify legitimate transactions based on the second set of features, and wherein the third set of one or more transaction fraud determinations comprises at least one machine learning model trained to identify fraudulent transactions based on the third set of features.

### Claim 5 (depends on 4)

The method of claim 4, wherein the set of static rule fraud determinations comprise one or more of: a validity determination of an application programming interface (API) key received with the transaction, a determination that the API key is associated with a merchant from which the transaction was received, a transaction type determination of the received transaction, a coherency check on payment information received with the transaction, an amount of the transaction is non-negative and/or valid, the parameters and/or API request is valid for the received transaction, and a check of the transaction location against a set of blocked transaction locations.

### Claim 6 (depends on 2)

The method of claim 2, wherein the third set of features comprises at least an order of magnitude number of features greater than a number of features in the second set of features, and wherein at least one feature in the third set of features is generated from a greater number of database records corresponding to the type of the features.

### Claim 7 (depends on 2)

The method of claim 2, further comprising:
in response to initiating performance of the first set of one or more transaction fraud determinations, the second set of one or more fraud determinations, or the third set of fraud determinations, generating an event message that requests a set of transaction features;
transmitting event message to a feature manager system causing the feature manager system to generate the requested set of features;
caching the generated requested set of features in a feature cache for a life of the transaction; and
clearing the generated requested set of features from the feature cache when the transaction is concluded.

### Claim 8 (depends on 7)

The method of claim 7, further comprising:
in response to receiving a request to generate the requested set of features, determining whether any feature in the requested set of features is partially or fully computed and stored in the feature cache;
for each fully computed feature stored in the feature cache, using the stored version of said each feature for the generation of the requested set of features;
for each partially computed feature stored in the feature cache, computing an update to a cached version of said each partially computed feature to transform said each partially computed feature into a fully computed feature, and caching the fully computed feature in the feature cache; and
for each new feature in the requested set of features that is not at least partially computed, computing said each new feature for the requested set of features, and caching said each new feature in the feature cache.

### Claim 9 (depends on 1)

The method of claim 1, further comprising:
processing or rejecting the transaction based on a result that identifies the transaction as fraudulent or legitimate.

### Claim 10 (depends on 1)

The method of claim 1, wherein the transaction attributes comprise one or more of a user identifier (ID), a card number, an email address, a telephone number, and a cookie ID, and wherein the transaction data comprises one or more of a dollar amount, a transaction date and/or time, a merchant ID for a merchant the generated the transaction request, an internet protocol address, a customer name, a shipping address, a billing address, and a purchase location.

### Claim 11

A non-transitory computer readable storage medium having instructions stored therein, which when executed by a server computer system, causes the server computer system to perform operations for fraud detection, the operations comprising:
receiving a transaction from a user having transaction attributes and transaction data;
initiating a progressive fraud determination process that determines whether the transaction is fraudulent or legitimate;
performing a first set of one or more transaction fraud determinations based on a first set of features determined for the transaction;
in response to a determination that the first set of one or more transaction fraud determinations has identified the transaction as fraudulent or legitimate, terminating the progressive fraud determination process;
in response to a determination that the first set of one or more transaction fraud determinations is inconclusive of determining whether the transaction is fraudulent or legitimate, performing a second set of one or more transaction fraud determinations based on a second set of features determined for the transaction, wherein the second set of features is greater than the first set of features; and
terminating the progressive fraud determination process based on a fraud determination rendered by the second set of one or more transaction fraud determinations that identifies the transaction as fraudulent or legitimate.

### Claim 12 (depends on 11)

The non-transitory computer readable storage medium of claim 11, wherein each of the second set of one or more transaction fraud determinations determines whether the transaction is legitimate, the operations further comprising:
in response to a determination that the second set of one or more transaction fraud determinations identify the transaction as legitimate, performing, by the server computer system, a third set of one or more transaction fraud determinations based on a third set of features determined for the transaction, wherein the third set of one or more transaction fraud determinations determines whether the transaction is fraudulent, and wherein the third set of features is a maximum set of transaction features that is greater than the second set of features; and
terminating, by the server computer system, the progressive fraud determination process to process or reject the transaction based on a fraud determination rendered by the third set of one or more transaction fraud determinations that identifies the transaction as fraudulent or legitimate.

### Claim 13 (depends on 12)

The non-transitory computer readable storage medium of claim 12, wherein terminating the progressive fraud determination process after the first set of one or more transaction fraud determinations, the second set of one or more transaction fraud determinations, or the third set of one or more transaction fraud determinations comprises:
transmitting an instruction to a transaction management system to process the transaction or reject the transaction based on a fraud determination result generated by the first set of one or more transaction fraud determinations or the second set of one or more transaction fraud determinations.

### Claim 14 (depends on 12)

The non-transitory computer readable storage medium of claim 12, wherein the first set of one or more transaction fraud determinations comprises a set of static rule fraud determinations, wherein the second set of one or more fraud determinations comprises at least one machine learning model trained to identify legitimate transactions based on the second set of features, and wherein the third set of one or more transaction fraud determinations comprises at least one machine learning model trained to identify fraudulent transactions based on the third set of features.

### Claim 15 (depends on 12)

The non-transitory computer readable storage medium of claim 12, the operations further comprising:
in response to initiating performance of the first set of one or more transaction fraud determinations, the second set of one or more fraud determinations, or the third set of fraud determinations, generating an event message that requests a set of transaction features;
transmitting event message to a feature manager system causing the feature manager system to generate the requested set of features;
caching the generated requested set of features in a feature cache for a life of the transaction; and
clearing the generated requested set of features from the feature cache when the transaction is concluded.

### Claim 16

A server computer system, comprising:
one or more processors; and
a memory storing instruction, which when executed by the one or more processors, cause the server computer system to perform operations, comprising:
receiving a transaction from a user having transaction attributes and transaction data;
initiating a progressive fraud determination process that determines whether the transaction is fraudulent or legitimate;
performing a first set of one or more transaction fraud determinations based on a first set of features determined for the transaction;
in response to a determination that the first set of one or more transaction fraud determinations has identified the transaction as fraudulent or legitimate, terminating the progressive fraud determination process;
in response to a determination that the first set of one or more transaction fraud determinations is inconclusive of determining whether the transaction is fraudulent or legitimate, performing a second set of one or more transaction fraud determinations based on a second set of features determined for the transaction, wherein the second set of features is greater than the first set of features; and
terminating the progressive fraud determination process based on a fraud determination rendered by the second set of one or more transaction fraud determinations that identifies the transaction as fraudulent or legitimate.

### Claim 17 (depends on 16)

The non-transitory computer readable storage medium of claim 16, wherein each of the second set of one or more transaction fraud determinations determines whether the transaction is legitimate, the operations further comprising:
in response to a determination that the second set of one or more transaction fraud determinations identify the transaction as legitimate, performing, by the server computer system, a third set of one or more transaction fraud determinations based on a third set of features determined for the transaction, wherein the third set of one or more transaction fraud determinations determines whether the transaction is fraudulent, and wherein the third set of features is a maximum set of transaction features that is greater than the second set of features; and
terminating, by the server computer system, the progressive fraud determination process to process or reject the transaction based on a fraud determination rendered by the third set of one or more transaction fraud determinations that identifies the transaction as fraudulent or legitimate.

### Claim 18 (depends on 17)

The non-transitory computer readable storage medium of claim 17, wherein the operations for terminating the progressive fraud determination process after the first set of one or more transaction fraud determinations, the second set of one or more transaction fraud determinations, or the third set of one or more transaction fraud determinations comprises:
transmitting an instruction to a transaction management system to process the transaction or reject the transaction based on a fraud determination result generated by the first set of one or more transaction fraud determinations or the second set of one or more transaction fraud determinations.

### Claim 19 (depends on 17)

The non-transitory computer readable storage medium of claim 17, wherein the first set of one or more transaction fraud determinations comprises a set of static rule fraud determinations, wherein the second set of one or more fraud determinations comprises at least one machine learning model trained to identify legitimate transactions based on the second set of features, and wherein the third set of one or more transaction fraud determinations comprises at least one machine learning model trained to identify fraudulent transactions based on the third set of features.

### Claim 20 (depends on 17)

The non-transitory computer readable storage medium of claim 17, the operations further comprising:
in response to initiating performance of the first set of one or more transaction fraud determinations, the second set of one or more fraud determinations, or the third set of fraud determinations, generating an event message that requests a set of transaction features;
transmitting event message to a feature manager system causing the feature manager system to generate the requested set of features;
caching the generated requested set of features in a feature cache for a life of the transaction; and
clearing the generated requested set of features from the feature cache when the transaction is concluded.
