Systems and methods for artificial intelligence-based personalized purchase recommendations

Patent metadata

Field	Value
Publication number	US-11995701-B2
Application number	US-202117456776-A
Country	US
Kind code	B2
Filing date	Nov 29, 2021
Priority date	Mar 29, 2019
Publication date	May 28, 2024
Grant date	May 28, 2024

Abstract

According to certain aspects of the disclosure, a computer-implemented method may be used for determining one or more vehicle recommendations. The method may include receiving data pertaining to a user's internet browsing activity. The received data may be indicative of the user's automotive vehicle preferences. The method may include comparing the received data to a collection of stored vehicle qualities. The method also may include identifying, based on the received data and the comparison of the received data to the collection of stored vehicle qualities, a vehicle characteristic of interest to the user. Using the vehicle characteristic of interest, one or more vehicle recommendations may be determined. One or more vehicle recommendations may be communicated to the user.

First claim

What is claimed is:

1. A computer-implemented method for determining one or more vehicle recommendations, comprising:
receiving a user's internet browsing history, wherein the received internet browsing history is obtained via a plug-in application operating in conjunction with an internet browser operating on a user device associated with the user;
accessing a neural net that has been trained, based on a plurality of literature sources that pertain to vehicles and that are indicative of a plurality of population characteristics and a plurality of vehicle characteristics, to output one or more likely population characteristics indicated by input data pertaining to one or more webpages;
using the neural net, and based on the received internet browsing history as the input data, developing a profile of the user indicative of one or more population characteristics of the user;
iteratively receiving updates to the user's internet browser history and updating the profile using the neural net based on the updates;
accessing a library that includes a listing of a plurality of purchaser populations, each purchaser population in the plurality of purchaser populations respectively associated with (i) one or more population characteristics and (ii) one or more vehicle characteristics;
comparing the one or more population characteristics of the user to the respective population characteristics of the plurality of purchaser populations stored in the library to identify one or more purchaser populations sharing at least one of the one or more population characteristics of the user;
identifying one or more of the vehicle characteristics associated with the one or more identified purchaser populations as one or more vehicle characteristics of interest to the user;
using the one or more vehicle characteristics of interest to the user, determining one or more vehicle recommendations; and
causing the user device to display the one or more vehicle recommendations.

2. The computer-implemented method of claim 1 , wherein the received internet browsing history is obtained while the user is visiting a first webpage, and wherein the one or more vehicle recommendations is communicated to the user while the user is visiting a second webpage, wherein the second webpage is different than the first webpage.

3. The computer-implemented method of claim 1 , further comprising:
receiving, via the user device, a question from the user;
generating a response to the question based on the one or more population characteristics of the user; and
causing the user device to display the response.

4. The computer-implemented method of claim 3 , wherein the question from the user pertains to a specific vehicle.

5. The computer-implemented method of claim 1 , wherein the one or more vehicle recommendations includes information unavailable on a webpage the user is currently viewing at the time the one or more vehicle recommendations is displayed.

6. The computer-implemented method of claim 1 , wherein the display of the one or more vehicle recommendations is implemented via the plug-in application operating on the user device.

7. The computer-implemented method of claim 1 , wherein the plurality of population characteristics includes one or more of economic status, automotive enthusiasm level, family status, location, interests, or hobbies.

8. The computer-implemented method of claim 1 , wherein the vehicle characteristic includes at least one of safety information, handling information, climate control, entertainment features, audio system information, decorative features, or storage features.

9. The computer-implemented method of claim 1 , wherein the one or more vehicle recommendations includes a selectable link.

10. A computer-implemented method for determining one or more vehicle recommendations, comprising:
iteratively receiving, via an application operating in conjunction with an internet browser, respective data pertaining to a user's internet browsing activity in using the internet browser;
accessing a neural net that has been trained, based on a plurality of literature sources that pertain to vehicles and that are indicative of a plurality of population characteristics and a plurality of vehicle characteristics, to output one or more likely population characteristics indicated by input data pertaining to one or more webpages;
using the respective received data pertaining to the user's internet browsing activity as the input data for the neural net, generating a profile of the user indicative of one or more population characteristics of the user;
upon each successive iteration that respective data pertaining to the user's internet browsing activity is received, iteratively updating the profile of the user, using the neural net, based on the respective data pertaining to the user's internet browsing activity;
accessing a library that includes a listing of a plurality of purchaser populations, each purchaser population in the plurality of purchaser populations respectively associated with (i) one or more population characteristics and (ii) one or more vehicle characteristics;
comparing the one or more population characteristics of the user to the respective population characteristics of the plurality of purchaser populations stored in the library to identify one or more purchaser populations sharing at least one of the one or more population characteristics of the user;
identifying one or more of the vehicle characteristics associated with the one or more identified purchaser populations as one or more vehicle characteristics of interest to the user;
using the one or more vehicle characteristics of interest to the user, determining one or more vehicle recommendations; and
causing a user device associated with the user to display the one or more vehicle recommendations.

11. The computer-implemented method of claim 10 , wherein the one or more population characteristics includes one or more of economic status, automotive enthusiasm level, family status, location, interests, or hobbies.

12. The computer-implemented method of claim 10 , wherein the one or more vehicle characteristics includes at least one of safety information, handling information, climate control, entertainment features, audio system information, decorative features, or storage features.

13. The computer-implemented method of claim 10 , wherein the one or more vehicle recommendations includes a selectable link.

14. The computer-implemented method of claim 10 , wherein the one or more vehicle recommendations includes information unavailable on a webpage the user is currently viewing at the time the one or more vehicle recommendations is displayed.

15. The computer-implemented method of claim 10 , wherein the profile of the user includes one or more of demographic data, interests, priorities, or budget information for the user.

16. The computer-implemented method of claim 10 , wherein the profile of the user is based at least in part upon a webpage from the respective received data pertaining to the user's internet browsing activity that is not related directly to vehicles, but that is indicative of one or more of values, interests, demographics, or preferences of the user.

17. A computer-implemented method for determining one or more vehicle recommendations, comprising:
iteratively receiving, via an application operating in conjunction with an internet browser, respective data pertaining to a user's internet browsing activity in using the internet browser;
accessing a neural net that has been trained, based on a plurality of literature

Assignees

• Capital One Services Llc

Inventors

• Tang Qiaochu
• Wylie Stephen
• Price Micah
• Duckworth Staevan
• Awoyemi Olalekan
• Dagley Geoffrey
• Hoover Jason

Classifications

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06Q30/0631 Primary

  Recommending goods or services · CPC title

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title

• G06N3/08

  Learning methods · CPC title