Systems and methods for providing improved recommendations

Patent metadata

Field	Value
Publication number	US-12511667-B2
Application number	US-202318526708-A
Country	US
Kind code	B2
Filing date	Dec 1, 2023
Priority date	Jul 13, 2018
Publication date	Dec 30, 2025
Grant date	Dec 30, 2025

Abstract

Systems and methods for providing improved recommendations are disclosed. In some embodiments, the systems and methods may be used for vehicle recommendations. The system may include a server system configured to receive user historical vehicle preferences, user vehicle preferences, generate weighted feature data sets, and apply a similarity model to the generated weighted feature data set in order to determine a vehicle recommendation data set. A visual representation of the vehicle recommendation data set may then be provided to an interface associated with a user.

First claim

We claim:

1 . A system for providing improved vehicle recommendations, the system comprising:
a processor; non-volatile memory; and a set of instructions on the non-volatile memory that are executable by the processor, configured for:
a user vehicle preference data input module configured to receive responses entered via a questionnaire regarding vehicles display by the user vehicle preference data input module via a user interface on a user device;
a user historical vehicle preference data input module configured to generate a user historical vehicle preference data set based on an analysis of browsing history on the user interface;
a vector generation module configured to map the responses, entered via the questionnaire regarding vehicles, in n-dimensional vehicle feature space to a plurality of vehicle features to generate a first feature vector comprising the plurality of vehicle features;
a similarity cosine module configured to determine a vehicle recommendation data set, comprising at least one vehicle and its associated vehicle features, by minimizing a distance between the first feature vector and a plurality of second vectors representing a plurality of vehicles; and
providing the vehicle recommendation data set to the user device.

2 . The system of claim 1 , wherein the analysis of the browsing history is based on other users' browsing history, and wherein the other users match a demographic in a demographic information of a user.

3 . The system of claim 1 , wherein the responses comprise a plurality of data units corresponding to the plurality of vehicle features.

4 . The system of claim 3 , wherein the plurality of vehicle features comprise a make, model, mileage, location, optional features, and condition.

5 . The system of claim 1 , wherein the set of instructions are further configured for a filter module configured to filter the vehicle recommendation data set based on one or more predetermined criteria.

6 . The system of claim 5 , wherein the one or more predetermined criteria comprise a maximum price.

7 . The system of claim 5 , wherein the one or more predetermined criteria comprise a requirement for each vehicle in the vehicle recommendation data set to comprise a unique combination of make and model.

8 . The system of claim 1 , wherein the first feature vector defines a first vector space and the user historical vehicle preference data set defines a second vector space, and wherein the first vector space and the second vector space have a different dimensionality.

9 . The system of claim 8 , wherein the set of instructions are further configured for:
a mapping module configured to map the second vector space to the first vector space based on attributes of a predetermined threshold of vehicles viewed in the browsing history.

10 . The system of claim 1 , wherein the responses correspond to a field for a vehicle feature.

11 . A method for providing improved vehicle recommendations comprising:
receiving, by a user vehicle preference data input module, responses entered from a questionnaire regarding vehicles displayed via a user interface on a user device; determining, by a user historical vehicle preference data input module, user historical vehicle preference data based on an analysis of browsing history on the user interface; generating a first feature vector, comprising a plurality of vehicle features, based on mapping the responses, entered via the questionnaire regarding vehicles, in n-dimensional vehicle feature space to the plurality of vehicle features; weighting the first feature vector based on the user historical vehicle preference data; determining, by a similarity cosine module, a vehicle recommendation data set, comprising at least one vehicle and its associated vehicle features, by minimizing a distance between the first feature vector and a plurality of second vectors representing a plurality of vehicles; and providing the vehicle recommendation data set to the user device.

12 . The method of claim 11 , further comprising receiving demographic information of a user; and
wherein the analysis of browsing history is based on other users' browsing history, and wherein the other users match a demographic in the demographic information of the user.

13 . The method of claim 11 , wherein the responses comprise a plurality of data units corresponding to the plurality of vehicle features.

14 . The method of claim 13 , wherein the plurality of vehicle features comprises a make, model, mileage, location, optional features, and condition.

15 . The method of claim 11 , further comprising filtering, by a filter module, the vehicle recommendation data set based on one or more predetermined criteria.

16 . The method of claim 15 , wherein the one or more predetermined criteria comprise a maximum price.

17 . The method of claim 15 , wherein the one or more predetermined criteria comprise a requirement for each vehicle in the vehicle recommendation data set to comprise a unique combination of make and model.

18 . The method of claim 11 , wherein the first feature vector defines a first vector space and the user historical vehicle preference data defines a second vector space, and wherein the first vector space and the second vector space have a different dimensionality.

19 . The method of claim 18 , further comprising:
mapping, by a mapping module, the second vector space to the first vector space based on attributes of a predetermined threshold of vehicles viewed in the browsing history.

20 . One or more non-transitory computer-readable medium storing a set of executable instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
receiving, by a user vehicle preference data input module, responses entered from a questionnaire regarding vehicles displayed via a user interface on a user device; determining, by a user historical vehicle preference data input module, user historical vehicle preference data based on an analysis of browsing history on the user interface; generating a first feature vector, comprising a plurality of vehicle features, based on mapping the responses, entered via the questionnaire regarding vehicles, in n-dimensional vehicle space to the plurality of vehicle features; weighting the first feature vector based on the user historical vehicle preference data; determining a vehicle recommendation data set, comprising at least one vehicle and its associated vehicle features, by minimizing a distance between a first vector and a plurality of second vectors representing a plurality of vehicles; and providing the vehicle recommendation data set to the user device.

Assignees

• Capital One Services Llc

Inventors

• Anderson Jacob
• Venuraju Mithra Kosur
• Mittal Shilpa
• Hoang Ben
• Deshpande Amit
• Shen Jie

Classifications

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G06Q30/0629

  by pre-processing results, e.g. ranking or ordering results · CPC title

• G06F16/24578

  using ranking · CPC title

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title

• G06F16/84

  Mapping; Conversion · CPC title