Fraud detection and transaction review

Patent metadata

Field	Value
Publication number	US-10062078-B1
Application number	US-201615181962-A
Country	US
Kind code	B1
Filing date	Jun 14, 2016
Priority date	Jun 14, 2016
Publication date	Aug 28, 2018
Grant date	Aug 28, 2018

Abstract

An automated purchase transaction service implements a two-phase analysis to identify suspect transactions and to freeze merchant accounts associated with certain of the suspect transactions. In a first analysis phase, a transaction is analyzed using a first predictive model to determine a probability that the transaction is fraudulent. If the probability exceeds a first threshold, the transaction is further analyzed in a second phase. In the second analysis phase, the transaction is analyzed using a second predictive model to determine a probability that manual review by a human analyst will result in freezing the associated account. If the probability exceeds a second threshold, the transaction is automatically frozen. If the probability does not exceed the second threshold, the transaction is submitted to a human analyst for manual review.

First claim

What is claimed is:

1. A method performed by one or more computers of a transaction processing system, the method comprising:
processing multiple purchase transactions between merchants and customers using point-of-sale (POS) computing devices, wherein each purchase transaction has a risk of being fraudulent;
receiving transaction data at the one or more computers of the transaction processing system, the transaction data being associated with the multiple purchase transactions, at least a portion of the transaction data being received from the POS computing devices;
storing the received transaction data as historical transaction data;
compiling training data, the training data comprising (a) the stored historical transaction data and (b) an indication, for each historical purchase transaction of multiple historical purchase transactions, of whether a human analyst froze a first merchant account as a result of analyzing historical transaction data associated with the historical purchase transaction;
creating a predictive model based at least in part on the training data, the predictive model producing a probability, based at least in part on given transaction data associated with a given transaction, that human analysis of the given transaction data will result in freezing a second merchant account;
receiving current transaction data for a current purchase transaction from a POS device associated with a third merchant account;
generating a first probability, based at least in part on the current transaction data, that the current purchase transaction is fraudulent;
determining that the first probability is greater than a first threshold;
in response to determining that the first probability is greater than the first threshold, generating a second probability, based at least in part on the current transaction data and the predictive model, that human analysis of the current transaction data will result in freezing the third merchant account; and
if the second probability is greater than a second threshold, freezing the third merchant account.

2. The method of claim 1 , further comprising:
if the second probability is not greater than the second threshold, initiating human analysis of the current transaction data to determine whether to freeze the third merchant account.

3. The method of claim 1 , further comprising:
generating first probabilities corresponding respectively to the multiple purchase transactions based at least in part on the transaction data associated with the multiple purchase transactions;
identifying a set of the multiple purchase transactions having corresponding first probabilities that are less than the first threshold and greater than a third threshold;
selecting a subset of the set of multiple purchase transactions, wherein the subset is less than all of the multiple purchase transactions of the set;
generating second probabilities for the purchase transactions of the subset;
for each purchase transaction of the subset, comparing the second probability of the purchase transaction with the second threshold to determine whether to initiate human analysis of the purchase transaction.

4. The method of claim 1 , further comprising:
generating second probabilities corresponding respectively to the multiple purchase transactions based at least in part on the predictive model and the transaction data of the multiple purchase transactions;
identifying a set of purchase transactions comprising less than all of the multiple purchase transactions having corresponding second probabilities that are greater than the second threshold; and
initiating human review of the purchase transactions of the set.

5. The method of claim 1 , further comprising:
analyzing the transaction data to generate first probabilities that respective purchase transactions are fraudulent;
identifying a set of the multiple purchase transactions having corresponding first probabilities that are greater than a third threshold, wherein the third threshold is greater than the first threshold; and
freezing merchant accounts associated with the set of multiple purchase transactions.

6. A system, comprising:
one or more processors;
one or more non-transitory computer-readable media storing instructions executable by the one or more processors, wherein the instructions program the one or more processors to perform actions comprising:
receiving transaction data associated with multiple purchase transactions, at least a portion of the transaction data being received from point-of-sale (POS) devices, the transaction data including first transaction data for a first purchase transaction;
generating a first probability, based at least in part on the first transaction data, that the first purchase transaction is fraudulent;
determining that the first probability is greater than a first threshold;
in response to determining that the first probability is greater than the first threshold, generating a second probability, based at least in part on the first transaction data and a first predictive model, that human analysis of the first transaction data will result in freezing a first account associated with the first transaction; and
if the second probability is greater than the second threshold, freezing the first account.

7. The system of claim 6 , wherein generating the first probability is based at least in part on a second predictive model.

8. The system of claim 6 , the actions further comprising:
if the second probability is not greater than the second threshold, initiating human analysis to determine whether to freeze the first account.

9. The system of claim 6 , the actions further comprising:
compiling training data, the training data comprising (a) at least a portion of the received transaction data; and (b) an indication, for each historical purchase transaction of multiple historical purchase transactions, of whether a human analyst froze a corresponding account as a result of analyzing the transaction data corresponding to the historical purchase transaction; and
creating the first predictive model based at least in part on the training data, the first predictive model producing a probability, based at least in part on given transaction data associated with a given transaction, that human analysis of the given transaction data will result in freezing an account associated with the given transaction.

10. The system of claim 6 , the actions further comprising:
generating first probabilities corresponding respectively to the multiple purchase transactions based at least in part on the transaction data associated with the multiple purchase transactions;
identifying a set of the multiple purchase transactions having corresponding first probabilities that are less than the first threshold and greater than a third threshold; and
selecting a subset of the set of multiple purchase transactions, wherein the subset is less than all of the multiple purchase transactions of the set;
generating second probabilities for the purchase transactions of the subset; and
for each purchase transaction of the subset, comparing the second probability of the purchase transaction with the second threshold to determine whether to initiate human analysis of the purchase transaction.

11. The system of claim 6 , the actions further comprising:
generating second probabilities corresponding respectively to the purchase transactions based at least in part on the predictive model and the transaction data associated with the purchase transactions;
identifying a set of the purchase transactions having corresponding second probabilities that are less than the second threshold and greater than a third thre

Assignees

• Square Inc

Inventors

• Boates Brian
• Zhou Rensheng
• Dalis Melissa

Classifications

• G06Q20/10

  specially adapted for electronic funds transfer [EFT] systems; specially adapted for home banking systems · CPC title

• G06Q20/4016 Primary

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

• G06Q20/209

  Specified transaction journal output feature, e.g. printed receipt or voice output · CPC title

• G06Q20/407

  Cancellation of a transaction · CPC title