Methods and systems for predicting account-level risk scores of cardholders

Patent metadata

Field	Value
Publication number	US-12217263-B2
Application number	US-202217739045-A
Country	US
Kind code	B2
Filing date	May 6, 2022
Priority date	May 8, 2021
Publication date	Feb 4, 2025
Grant date	Feb 4, 2025

Abstract

Embodiments provide artificial intelligence-based methods and systems for predicting account-level risk scores associated with cardholders. Method performed by server system includes accessing payment transaction data and cardholder risk data associated with cardholder. The payment transaction data includes transaction variables associated with past payment transactions performed at Point of Interaction (POI) terminals within a particular time window. Method includes generating cardholder profile data based on the transaction variables and the cardholder risk data. Method includes determining account-level risk scores associated with the cardholder based on cardholder profile data. Each account-level risk score of account-level risk scores is determined by a trained machine learning model. The account-level risk scores include a wallet reload risk score, an account reissuance risk score, and a transaction channel risk score. Further, the method includes transmitting a recommendation message to an issuer server associated with the cardholder based on the account-level risk scores.

First claim

We claim:

1. A computer-implemented method, comprising:
accessing, by a server system from a transaction database via a network, payment transaction data and cardholder risk data associated with a cardholder, the payment transaction data comprising a plurality of transaction variables associated with past payment transactions performed at one or more Point of Interaction (POI) terminals within a particular time window;
generating, by the server system, cardholder profile data associated with the cardholder based, at least in part, on the plurality of transaction variables and the cardholder risk data, the cardholder profile data comprising: (a) wallet transaction features associated with a set of crypto-wallets, (b) spend behavioral pattern features of the cardholder for a plurality of merchant terminals associated with a plurality of merchants, and (c) cardholder transaction features according to merchant risk profiles;
determining, by the server system, a plurality of account-level risk scores associated with the cardholder based, at least in part, on the cardholder profile data, the plurality of account-level risk scores determined by a plurality of trained machine learning models, the plurality of account-level risk scores comprising: (a) a wallet reload risk score, (b) an account reissuance risk score, and (c) a transaction channel risk score, the plurality of trained machine learning models comprising: (a) a wallet reload risk model, (b) an account reissuance risk model, and (c) a transaction channel risk model;
wherein determining the plurality of account-level risk scores comprises:
generating, by the server system using the wallet reload risk model, the wallet reload risk score associated with the cardholder and a crypto-wallet, the wallet reload risk score indicating a likelihood of the cardholder performing risky wallet transactions with the crypto-wallet in a subsequent time interval based, at least in part, on the wallet transaction features associated with the set of crypto-wallets;
generating, by the server system using the account reissuance risk model, the account reissuance risk score based, at least in part, on the spend behavioral pattern features of the cardholder, the account reissuance risk score indicating a probability of one or more payment cards of the cardholder to be reissued within the subsequent time interval, wherein the spend behavioral pattern features comprise one or more of: (a) merchant features based on common point of purchase (CPP) compromise risk scores of the plurality of merchants, (b) cardholder interaction features with the plurality of merchants classified into a set of merchant classes, (c) cardholder risk features, or (d) payment network related risk features; and
generating, by the server system using the transaction channel risk model, the transaction channel risk score associated with the cardholder based, at least in part, on the merchant risk profiles including one or more of a location and a processor of a risky merchant where a payment card number was stolen, the transaction channel risk score indicating a probability of the one or more payment cards of the cardholder being compromised while transacting with the risky merchant within a subsequent time period; and

transmitting, via the network by a recommendation engine associated with the server system, a recommendation message to an issuer server associated with the cardholder based, at least in part, on the plurality of account-level risk scores for performing an action in response to the recommendation message.

2. The computer-implemented method as claimed in claim 1 , wherein the plurality of transaction variables comprises at least: POI terminal type indicator, payment transaction type indicator, a number of spend transactions at the one or more POI terminals within the particular time window, authorization data, and fraud risk data.

3. The computer-implemented method as claimed in claim 1 , wherein the one or more POI terminals comprise a set of crypto-wallets and merchant terminals.

4. The computer-implemented method as claimed in claim 1 , further comprising:
training the wallet reload risk model, the account reissuance risk model, and the transaction channel risk model using the payment transaction data stored in the transaction database.

5. The computer-implemented method as claimed in claim 1 , further comprising:
classifying, by the server system, a set of payment transactions of the cardholder into risky merchant transactions and non-risky merchant transactions based, at least in part, on the cardholder transaction features and the transaction channel risk model; and
generating, by the server system, the transaction channel risk score associated with the cardholder based, at least in part, on the classifying step.

6. The computer-implemented method as claimed in claim 1 , wherein the recommendation message is transmitted to the issuer server in real-time when the cardholder is performing a payment transaction or a crypto-wallet transaction.

7. The computer-implemented method as claimed in claim 1 , wherein the wallet reload risk model is a gradient boosting decision tree (GBDT) based classification model.

8. The computer-implemented method as claimed in claim 1 , wherein the action comprises: (a) accepting or declining reloading of the crypto-wallet, (b) authorizing or declining a payment transaction, or (c) reissuing a payment card to the cardholder.

9. A server system, comprising:
a communication interface;
a memory comprising machine-readable instructions; and
a processor communicably coupled to the communication interface and the memory, the processor configured to execute the machine-readable instructions to cause the server system at least in part to:
access, from a transaction database via a network, payment transaction data and cardholder risk data associated with a cardholder, the payment transaction data comprising a plurality of transaction variables associated with past payment transactions performed at one or more Point of Interaction (POI) terminals within a particular time window;
generate cardholder profile data associated with the cardholder based, at least in part, on the plurality of transaction variables and the cardholder risk data, the cardholder profile data comprising: (a) wallet transaction features associated with a set of crypto-wallets, (b) spend behavioral pattern features of the cardholder for a plurality of merchant terminals associated with a plurality of merchants, and (c) cardholder transaction features according to merchant risk profiles;
determine a plurality of account-level risk scores associated with the cardholder based, at least in part, on the cardholder profile data, the plurality of account-level risk scores determined by a plurality of trained machine learning models, the plurality of account-level risk scores comprising: (a) a wallet reload risk score, (b) an account reissuance risk score, and (c) a transaction channel risk score, the plurality of trained machine learning models comprising: (a) a wallet reload risk model, (b) an account reissuance risk model, and (c) a transaction channel risk model;

wherein determining the plurality of account-level risk scores comprises:
generating, using the wallet reload risk model, the wallet reload risk score associated with the cardholder and a crypto-wallet, the wallet reload risk score indicating a likelihood of the cardholder performing risky wallet transactions with the crypto-wallet in a subsequent time interval based, at least in part, on the wallet transaction features associated with the set of crypto-wallets;
generating, using the account reissuance risk model, the account reissuance risk score based, at least in part

Assignees

• Mastercard International Inc

Inventors

• Pandillapalli Bhargav
• Ranjan Rajesh Kumar
• Saraswat Ankur
• Gangwar Kshitij
• Kant Kamal
• Syngal Sonali
• Powar Suhas
• Das Debasmita
• Nath Pritam Kumar
• Pant Nishant
• Katyal Yatin
• Kumar Nitish
• Singh Karamjit

Classifications

• G06Q40/03 Primary

  Credit; Loans; Processing thereof · CPC title

• G06Q20/389

  Keeping log of transactions for guaranteeing non-repudiation of a transaction · CPC title

• G06Q20/36

  using electronic wallets or electronic money safes · CPC title

• G06Q20/4016 Primary

  involving fraud or risk level assessment in transaction processing · CPC title