System and method for account transaction and balance prediction

Patent metadata

Field	Value
Publication number	US-11328366-B2
Application number	US-202016895228-A
Country	US
Kind code	B2
Filing date	Jun 8, 2020
Priority date	Jan 21, 2014
Publication date	May 10, 2022
Grant date	May 10, 2022

Abstract

A system includes a financial data storage that stores internal account data, external account data, and transaction data, a communications interface configured to receive external account data from an external data source and transaction data from a merchant, and a database manager that cooperates with the communications interface to store the external account data and the transaction data. The system also includes an account balance prediction processor configured to identify recurring expense transaction(s) and recurring income transaction(s) and update them based on user input, determine predicted expense transaction(s) and a predictive income using a predictive model, and calculate a predicted account balance. The system also includes a transaction monitor configured to compare one or more actual expense transactions with one or more predicted expense transactions, and an alert transmitter configure to generate and transmit one or more alerts based on the comparison.

First claim

What is claimed is:

1. A method comprising:
identifying one or more expense transactions and one or more income transactions based on data obtained from a financial data storage associated with an account holder;
receiving, from a mobile device associated with the account holder, a user input identifying the one or more expense transactions or the one or more income transactions as either recurring transactions or one-time transactions;
classifying the one or more expense transactions or the one or more income transactions based on the user input;
determining one or more predicted expense transactions using machine learning algorithms based on the one or more recurring expense transactions and a prediction time period;
determining a predicted income using the machine learning algorithms based on the one or more recurring income transactions and the prediction time period;
receiving, at a communications interface, one or more actual expense transactions within the predicted time period;
comparing the one or more actual expense transactions with the one or more predicted expense transactions;
transmitting, in real time via the communications interface, a first alert via a network when a first predicted expense transaction of the one or more predicted expense transactions does not match the one or more actual expense transactions based on one or more fuzzy matching algorithms; and
transmitting a second alert when the one or more predicted expense transactions exceed the predicted income.

2. The method of claim 1 , wherein the one or more actual expense transactions are received from a third party that bills the account holder on a recurring basis.

3. The method of claim 1 , further comprising:
providing, with a user interface on the mobile device, a predicted account balance to the account holder based on the predicted income and the one or more predicted expense transactions.

4. The method of claim 1 , wherein the financial data storage and the communications interface are integrated into a backend system of a financial service provider.

5. The method of claim 1 , wherein the data obtained from the financial storage comprises bill pay data of the account holder.

6. The method of claim 1 , further comprising:
evaluating the data obtained from the financial storage by applying seasonal, demographic, and geographic trend factors to calculate a volatility factor of credits and debits associated with the account holder that provides confidence associated with the one or more predicted expense transactions and the predicted income.

7. The method of claim 6 , further comprising:
providing a range associated with each of the predicted income and the one or more predicted expense transactions to illustrate a first end prediction and a second end prediction associated with each, the range based on the volatility factor.

8. A method comprising:
receiving account holder data comprising internal account data, external account data, and transaction data associated with an account holder;
identifying one or more recurring expense transactions based on the account holder data;
verifying the one or more recurring expense transactions based on a user input;
determining one or more predicted expense transactions based on the one or more recurring expense transactions and a prediction time period;
receiving one or more actual expense transactions from an external data source within the prediction time period;
comparing the one or more actual expense transactions with the one or more predicted expense transactions; and
transmitting, to a mobile device associated with the account holder, an alert when a first predicted expense transaction has not been matched with the one or more actual expense transactions within a predetermined time period, the comparison based on one or more fuzzy matching algorithms.

9. The method of claim 8 , further comprising:
determining a predicted account balance based on the one or more predicted expense transactions and a predicted income, the predicted income based on one or more recurring income transactions and the prediction time period.

10. The method of claim 9 , further comprising:
evaluating the internal account data, the external account data and the transaction data by applying seasonal, demographic, and geographic trend factors to calculate a volatility factor of credits and debits associated with the account holder that provides confidence associated with the one or more predicted expense transactions, the predicted income, and the predicted account balance.

11. The method of claim 10 , further comprising:
providing a range associated with each of the predicted account balance, the one or more predicted expense transactions, and the predicted income to illustrate a first end prediction and a second end prediction associated with each, the range based on the volatility factor.

12. The method of claim 8 , further comprising:
providing, with a user interface on the mobile device, a user interface displaying the one or more predicted expense transactions to the account holder.

13. A system comprising:
one or more processors; and
a memory in communication with the one or more processors and storing instructions that, when executed by the one or more processors, are configured to cause the system to:
receive account holder data from a financial data storage, the account holder data associated with an account holder;
identify one or more recurring expense transactions;
determine one or more predicted expense transactions based on the one or more recurring expense transactions and a prediction time period;
receive one or more actual expense transactions within the prediction time period;
compare the one or more actual expense transactions with the one or more predicted expense transactions;
determine that a difference between the one or more actual expense transactions and the one or more predicted expense transactions is greater than a predefined threshold; and
transmit, to a mobile device associated with the account holder, a first alert indicating that the difference between the one or more actual expense transactions and the one or more predicted expense transactions exceeds the predefined threshold.

14. The system of claim 13 , wherein the instructions, when executed by the one or more processors, are further configured to cause the system to receive, from the mobile device associated with the account holder, a user input identifying any of the one or more recurring expense transactions that are one-time transactions.

15. The system of claim 13 , wherein the one or more actual expense transactions are received from a third party that bills the account holder on a recurring basis.

16. The system of claim 13 , wherein the instructions, when executed by the one or more processors, are further configured to cause the system to:
provide, on the mobile device, a user interface displaying the first alert and the one or more predicted expense transactions to the account holder.

17. The system of claim 13 , wherein the one or more processors and the memory are integrated into a backend system of a financial service provider.

18. The system of claim 13 , wherein the one or more recurring transactions comprises bill pay data of the account holder.

19. The system of claim 13 , wherein the instructions, when executed by the one or more processors, are further configured to cause the system to:
evaluate the account holder data by applying seasonal, demographic, and geographic trend factor

Assignees

• Capital One Services Llc

Inventors

• Rephlo Jeremy
• Struble Daniel L
• Lafeer Mohamed
• Rowley Robert D
• Naqui Nondini A
• Calandro Sarah
• Hart Colin

Classifications

• G06N20/00

  Machine learning · CPC title

• G06Q40/12 Primary

  Accounting · CPC title