Vehicle recommendations based on browser context

Patent metadata

Field	Value
Publication number	US-2024273599-A1
Application number	US-202318169510-A
Country	US
Kind code	A1
Filing date	Feb 15, 2023
Priority date	Feb 15, 2023
Publication date	Aug 15, 2024
Grant date	—

Abstract

In some implementations, a recommendation system may receive information related to a browser context associated with the client device. The recommendation system may generate a vehicle feature vector that includes an array of elements to represent a plurality of vehicle attributes. The recommendation system may apply a similarity model to the vehicle feature vector to determine a vehicle recommendation dataset that includes a plurality of vehicles that are each associated with a respective set of vehicle attributes. The recommendation system may filter the vehicle recommendation dataset based on a subset of the information related to the browser context associated with the client device that indicates a profile of a user associated with the client device. The recommendation system may provide information related to the vehicle recommendation dataset to the client device for display in an interface of the client device.

First claim

What is claimed is:

1 . A system for providing vehicle recommendations based on browser context, the system comprising:
one or more memories; and one or more processors, communicatively coupled to the one or more memories, configured to:
receive, from an application executing on a client device, information related to a browser context associated with the client device,
wherein the browser context includes a historical pattern of browser interactions with content related to vehicles;

generate a vehicle feature vector that includes an array of elements to represent a plurality of vehicle attributes,
wherein each element in the array of elements has a value to represent a user preference related to a corresponding vehicle attribute based on the historical pattern of browser interactions with the content related to vehicles;

apply a similarity model to the vehicle feature vector to determine a vehicle recommendation dataset that includes a plurality of vehicles that are each associated with a respective set of vehicle attributes;
filter the vehicle recommendation dataset based on a subset of the information related to the browser context associated with the client device that indicates one or more financing preferences associated with a user associated with the client device; and
provide information related to the vehicle recommendation dataset to the client device for display in an interface associated with the client device.

2 . The system of claim 1 , wherein the one or more processors are further configured to:
determine a location associated with the client device based on the information related to the browser context associated with the client device,
wherein the vehicle recommendation dataset is filtered such that the information displayed in the interface associated with the client device is related to vehicles that are within a threshold distance of the location associated with the client device.

3 . The system of claim 1 , wherein the one or more processors are further configured to:
store vectorized information associated with each vehicle in a vehicle inventory; and determine the vehicle recommendation dataset based on respective cosine distances between the vehicle feature vector and the vectorized information associated with each vehicle in the vehicle inventory.

4 . The system of claim 1 , wherein the one or more processors, to filter the vehicle recommendation dataset, are configured to:
determine, for each of the plurality of vehicles included in the vehicle recommendation dataset and based on the one or more financing preferences associated with the user associated with the client device, a probability that the user will qualify to enter into a transaction associated with the vehicle; and remove, from the plurality of vehicles included in the vehicle recommendation dataset, one or more vehicles based on the probability that the user will qualify to enter into a transaction for each of the plurality of vehicles included in the vehicle recommendation dataset.

5 . The system of claim 1 , wherein the historical pattern of browser interactions includes one or more browser sessions in which the client device accessed a vehicle manufacturer or vehicle dealer website to view or configure a vehicle having a set of vehicle attributes.

6 . The system of claim 1 , wherein the browser context includes a browser fingerprint that is based on the historical pattern of browser interactions with content related to vehicles, metadata obtained from one or more services that indicates the one or more financing preferences associated with the user, and a profile associated with advertising content presented to the user.

7 . The system of claim 1 , wherein the one or more processors are configured to:
receive, from the application executing on the client device, information related to a browser interaction with content related to a vehicle; and determine, based on the one or more financing preferences associated with the user associated with the client device, that a probability of the user qualifying to enter into a transaction for the vehicle associated with the browser interaction fails to satisfy a threshold,
wherein the information related to the vehicle recommendation dataset indicates a vehicle associated with one or more of a different year, make, or model than the vehicle associated with the browser interaction based on a probability of the user qualifying to enter into a transaction for the vehicle associated with the different year, make, or model satisfying the threshold.

8 . The system of claim 1 , wherein the one or more processors are configured to:
apply a filter to the vehicle recommendation dataset to ensure that each vehicle in the vehicle recommendation set has a different vehicle make and vehicle model combination.

9 . The system of claim 1 , wherein the plurality of vehicle attributes includes one or more of a year, a make, a model, a mileage, a price, a fuel efficiency, an engine or a motor, a fuel type, a drive train, an exterior color, a body style, a condition, or a transmission.

10 . A method for providing context-aware recommendations, comprising:
receiving, by a device from an application executing on a client device, information related to a browser context associated with the client device,
wherein the browser context includes a historical pattern of browser interactions with content related to vehicles and indicates a location associated with the client device;
generating, by the device, a vehicle feature vector that includes an array of elements to represent a plurality of vehicle attributes,
wherein each element in the array of elements has a value to represent a user preference related to a corresponding vehicle attribute based on the historical pattern of browser interactions with the content related to vehicles;
applying, by the device, a similarity model to the vehicle feature vector to determine a vehicle recommendation dataset that includes a plurality of vehicles that are each associated with a respective set of vehicle attributes; filtering, by the device, the vehicle recommendation dataset based on a subset of the information related to the browser context associated with the client device that indicates a profile of a user associated with the client device,
wherein the vehicle recommendation dataset is filtered such that the vehicle recommendation dataset excludes any vehicles that are not within a threshold distance of the location associated with the client device; and
providing, by the device, information related to the vehicle recommendation dataset to the client device for display in an interface associated with the client device.

11 . The method of claim 10 , further comprising:
storing vectorized information associated with each vehicle in a vehicle inventory; and determining the vehicle recommendation dataset based on a cosine distance between the vehicle feature vector and the vectorized information associated with each vehicle in the vehicle inventory.

12 . The method of claim 10 , wherein filtering the vehicle recommendation dataset comprises:
determining, based on the profile of the user associated with the client device, a probability that the user will qualify to enter into a transaction for each of the plurality of vehicles included in the vehicle recommendation dataset; and removing, from the plurality of vehicles included in the vehicle recommendation dataset, one or more vehicles based on the probability that the user will qualify to enter into a transaction for each of the plurality of vehicles included in the vehicle recommendation dataset.

Assignees

• Capital One Services Llc

Inventors

• Danna Russell
• Nehmer Landon M
• Anderson Jacob
• Garner Michael

Classifications

• G06F16/9537

  Spatial or temporal dependent retrieval, e.g. spatiotemporal queries · CPC title

• G06Q30/0631 Primary

  Recommending goods or services · CPC title

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title