Utilizing machine learning and transaction data to determine fuel prices at fuel stations

Patent metadata

Field	Value
Publication number	US-12437312-B2
Application number	US-202318508392-A
Country	US
Kind code	B2
Filing date	Nov 14, 2023
Priority date	Nov 3, 2020
Publication date	Oct 7, 2025
Grant date	Oct 7, 2025

Abstract

A fuel price determination system may receive transaction data identifying purchases, made with transaction cards and mobile transaction card applications of client devices, of fuel at fuel stations. The fuel price determination system may receive location data identifying locations associated with users of the client devices and the transaction cards, and user history data associated with prior purchases of fuel at fuel stations by the users. The fuel price determination system may process the transaction data, location data, and user history data, with a machine learning model, to determine fuel prices at the fuel stations. The fuel price determination system may determine a ranked list of particular fuel stations in a geographical area based on the fuel prices and populate, based on the location data and the ranked list, a map to identify the particular fuel stations and the fuel prices at the particular fuel stations.

First claim

What is claimed is:

1. A method, comprising:
determining, by a device and based on sensor information associated with a vehicle, a first fuel amount and a second fuel amount;
processing, by the device, transaction data, location data, user history data, the first fuel amount, and the second fuel amount with a machine learning model, to determine a predicted fuel price for a predicted grade of fuel,
wherein the machine learning model is trained based on historical transaction data, historical location data, and historical user history data to determine the predicted fuel price, and
wherein processing the transaction data, the location data, the user history data, the first fuel amount, and the second fuel amount with the machine learning model comprises:
using machine learning to extrapolate the predicted fuel price from a fuel price for another grade of fuel when insufficient information exists for the predicted grade of fuel;

determining, by the device, a confidence score associated with the predicted fuel price based on the transaction data, the location data, or the user history data;
populating, by the device and based on the location data, a map to identify the predicted fuel price, an associated fuel station, and an indicator of the confidence score;
generating, by the device, a list of fuel stations based on the map;
automatically selecting, by the device, a fuel station from the generated list of fuel stations;
deploying, by the device and based on generating directions to the selected fuel station, an autonomous vehicle to the selected fuel station;
causing, by the device and based on deploying the autonomous vehicle to the selected fuel station, the autonomous vehicle to capture one or more images associated with the selected fuel station; and
retraining, by the device, the machine learning model based on the one or more images.

2. The method of claim 1 , wherein determining the confidence score comprises:
increasing or decreasing the confidence score based on determining whether a transaction amount associated with the predicted fuel price is associated with items other than fuel.

3. The method of claim 1 , wherein determining the confidence score comprises:
determining the confidence score based on a confidence of the predicted grade of fuel based on the user history data.

4. The method of claim 1 , further comprising:
determining whether to replace a previous fuel price with the predicted fuel price based on the confidence score satisfying a threshold.

5. The method of claim 1 , further comprising:
eliminating another predicted fuel price based on another confidence score, associated with the other predicted fuel price, failing to satisfy a threshold.

6. The method of claim 1 , wherein the indicator of the confidence score comprises a text indicator that indicates the predicted fuel price is estimated.

7. The method of claim 1 , further comprising:
extracting, from the one or more images, fuel price data associated with the selected fuel station;
updating, based on extracting the fuel price data associated with the selected fuel station, a predicted fuel price associated with the selected fuel station; and
updating the generated list of fuel stations based on the updated predicted fuel price associated with the selected fuel station.

8. A device, comprising:
one or more memories; and
one or more processors, coupled to the one or more memories, configured to:
determine, based on sensor information associated with a vehicle, a first fuel amount and a second fuel amount;
process transaction data, location data, user history data, the first fuel amount, and the second fuel amount with a machine learning model, to determine a predicted fuel price for a predicted grade of fuel,
wherein the machine learning model is trained based on historical transaction data, historical location data, and historical user history data to determine the predicted fuel price, and
wherein the one or more processors, to process the transaction data, the location data, the user history data, the first fuel amount, and the second fuel amount with the machine learning model, are configured to:
use machine learning to extrapolate the predicted fuel price from a fuel price for another grade of fuel when insufficient information exists for the predicted grade of fuel;

determine a confidence score associated with the predicted fuel price based on the transaction data, the location data, or the user history data;
populate, based on the location data, a map to identify the predicted fuel price, an associated fuel station, and an indicator of the confidence score;
generate a list of fuel stations based on the map;
automatically select a fuel station from the generated list of fuel stations;
deploy, based on generating directions associated with the selected fuel station, an autonomous vehicle to the selected fuel station;
cause, based on deploying the autonomous vehicle to the selected fuel station, the autonomous vehicle to capture one or more images associated with the selected fuel station; and
retrain the machine learning model based on the one or more images.

9. The device of claim 8 , wherein the one or more processors, to determine the confidence score, are configured to:
increase or decrease the confidence score based on determining whether a transaction amount associated with the predicted fuel price is associated with items other than fuel.

10. The device of claim 8 , wherein the one or more processors, to determine the confidence score, are configured to:
determine the confidence score based on a confidence of the predicted grade of fuel based on the user history data.

11. The device of claim 8 , wherein the one or more processors are further configured to:
determine whether to replace a previous fuel price with the predicted fuel price based on the confidence score satisfying a threshold.

12. The device of claim 8 , wherein the one or more processors are further configured to:
eliminate another predicted fuel price based on another confidence score, associated with the other predicted fuel price, failing to satisfy a threshold.

13. The device of claim 8 , wherein the indicator of the confidence score comprises a text indicator that indicates the predicted fuel price is estimated.

14. The device of claim 8 , wherein the one or more processors are further configured to:
provide an overlay identifying the predicted fuel price, the associated fuel station, and the indicator of the confidence score via a navigation application.

15. A non-transitory computer-readable medium storing a set of instructions, the set of instructions comprising:
one or more instructions that, when executed by one or more processors of a device, cause the device to:
determine, based on sensor information associated with a vehicle, a first fuel amount and a second fuel amount;
process transaction data, location data, user history data, the first fuel amount, and the second fuel amount with a machine learning model, to determine a predicted fuel price for a predicted grade of fuel,
wherein the machine learning model is trained based on historical transaction data, historical location data, and historical user history data to determine the predicted fuel price, and
wherein the one or more instructions, that cause the device to process the transaction data, the location data, the user history data, the first fuel amount, and the second fuel amount with the machine learning model, cause the device to:
use machine learning to extrapolate the predicted fuel price from a fuel price for another grade

Assignees

• Capital One Services Llc

Inventors

• ZARAKAS James
• Vukich Adam
• Johnson Molly

Classifications

• G06Q30/0203

  Market surveys; Market polls · CPC title

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G06F18/23

  Clustering techniques · CPC title

• G06Q50/06 Primary

  Energy or water supply · CPC title

• G06N20/00

  Machine learning · CPC title