# US 12,008,455 B2 · Personalized vehicle recommender system

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,008,455 B2
- Document ID: US12008455B2
- Status: Granted
- Kind: B2
- Application number: 16/444,513
- Filed: 2019-06-18
- Published: 2021-06-03
- Granted: 2024-06-11
- Earliest claimed: 2019-06-18
- Assignees: Driverbase Inc.
- Applicants: Driverbase Inc.
- Inventors: Dan Jackson; Andrew Bays; Eric Parziale
- Web page: https://patentgrep.com/patent/US12008455B2

## CPC

- G06N 3/042: Knowledge-based neural networks; Logical representations of neural networks
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/088: Non-supervised learning, e.g. competitive learning
- G06N 3/09: Supervised learning
- G06Q 30/0631: {Recommending goods or services}
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

A vehicle recommender system includes a storage system storing a trained neural network trained to generate a vehicle recommendation, a vehicle database comprising vehicle attribute data, and a driver database comprising driver profile data. A processing system communicates with the storage system and is configured to generate enriched driver profile data by comparing user-generated driver profile data to the location characteristic data and social characteristic data. The processing system calculates a driver attribute target value for each vehicle attribute category in a list of vehicle attribute categories based on the driver profile data, and then generates, using the trained neural network, at least one vehicle recommendation based on at least the vehicle attribute data and the driver attribute target values.

## Claims

### Claim 1

A vehicle recommender system comprising:
a storage system storing a trained neural network trained to generate a vehicle recommendation, wherein the neural network is trained on vehicle database data and driver database data;
a vehicle database comprising vehicle attribute data;
a driver database comprising user-generated driver profile data, wherein the user-generated driver profile data includes at least one driving location, at least one driving route, a driver age, a driver occupation, a current vehicle, a current vehicle ownership type, a past vehicle, and a past vehicle ownership type;
a processing system communicatively connected to the storage system and configured to:
generate enriched driver profile data by comparing the user-generated driver profile data to location characteristic data and/or to social characteristic data, wherein the enriched driver profile data includes an expected vehicle spend value and each of an average snowfall, an average altitude, road condition information, an average driving speed, and an estimated commute time;
calculate a driver attribute target value for each vehicle attribute category in a list of vehicle attribute categories based on the driver profile data, which includes the enriched driver profile data and the user-generated driver profile data; and
using the trained neural network, generate at least one vehicle recommendation based on at least the vehicle attribute data, the driver attribute target values and the driver profile data, including the enriched driver profile data and the user-generated driver profile data, wherein the trained neural network functions as a hybrid recommender system performing content-based filtering and collaborative filtering.

### Claim 2 (depends on 1)

The system of claim 1, wherein the location characteristic data includes data from the National Weather Service.

### Claim 3 (depends on 1)

The system of claim 1, wherein the social characteristic data includes census data.

### Claim 4 (depends on 1)

The system of claim 1, wherein the enriched driver profile data further includes an income value.

### Claim 5 (depends on 1)

The system of claim 1, wherein the user-generated driver profile data further includes at least one corresponding vehicle rating for each current vehicle and past vehicle.

### Claim 6 (depends on 1)

The system of claim 1, wherein the processing system is further configured to determine at least one vehicle purchase recommendation based on the at least one vehicle recommendation, wherein the vehicle purchase recommendation includes at least one of a recommended purchase location, a recommended inventory, a recommended purchase time, and a recommended purchase method.

### Claim 7 (depends on 1)

The system of claim 1, wherein the neural network is trained using a Weighted Approximate-Rank Pairwise (WARP) loss function.

### Claim 8

A method of generating a vehicle recommendation for a driver, the method comprising:
receiving driver profile data about a driver, wherein the driver profile data includes user-generated driver profile data and enriched profile data;
wherein the user-generated driver profile data includes at least one driving location, at least one driving route, a driver age, a driver occupation, a current vehicle, a current vehicle ownership type, a past vehicle, and a past vehicle ownership type;
wherein the enriched driver profile data includes an expected vehicle spend value and each of an average snowfall, an average altitude, road condition information, an average driving speed, and an estimated commute time;
receiving vehicle attribute data regarding each vehicle class in a list of vehicle classes;
receiving a list of vehicle attribute categories;
generating a driver attribute target value for each vehicle attribute category based on the driver profile data; and
providing the enriched driver profile data, the user-generated driver profile data, the vehicle attribute data, and the driver attribute target values as input data to a trained neural network to generate, by the trained neural network, at least one vehicle recommendation, wherein the trained neural network functions as a hybrid recommender system performing content-based filtering and collaborative filtering and wherein the neural network is trained on vehicle attribute data and driver profile data.

### Claim 9 (depends on 8)

The method of claim 8, further comprising:
receiving location characteristic data and social characteristic data;
further comprising generating the driver attribute target value based on the location characteristic data and the social characteristic data.

### Claim 10 (depends on 9)

The method of claim 9, wherein the location characteristic data includes data from the National Weather Service and wherein the social characteristic data includes census data.

### Claim 11 (depends on 8)

The method of claim 8, wherein the driver profile data further includes at least one of a current vehicle, a current vehicle ownership type, a past vehicle, and a past vehicle ownership type.

### Claim 12 (depends on 11)

The method of claim 11, wherein the user-generated driver profile data further includes at least one corresponding vehicle rating for each current vehicle and past vehicle.

### Claim 13 (depends on 8)

The method of claim 8, further comprising training the neural network trained using a Weighted Approximate-Rank Pairwise (WARP) loss function, and wherein the trained neural network functions as the hybrid recommender system performing the content-based filtering and the collaborative filtering of the driver profile data, the vehicle attribute data, and the driver attribute target values.

### Claim 14

A method of generating a vehicle recommendation for a driver, the method comprising:
storing a list of vehicle attribute categories;
storing vehicle attribute data for each vehicle class in a list of vehicle classes, wherein the vehicle attribute data includes, for each vehicle class, a vehicle attribute value for each vehicle attribute category;
storing user-generated driver profile data inputted by a driver, wherein the user-generated driver profile data includes at least one driving location, at least one driving route, a driver age, and a driver occupation, a current vehicle, a current vehicle ownership type, a past vehicle, a past vehicle ownership type, a current owned vehicle rating, and a past owned vehicle rating;
storing location characteristic data and social characteristic data;
generating enriched driver profile data by comparing the user-generated driver profile data to the location characteristic data and/or to the social characteristic data, wherein the enriched driver profile data includes an expected vehicle spend value and each of an average snowfall, an average altitude, road condition information, an average driving speed, and an estimated commute time;
calculating a driver attribute target value for each vehicle attribute category in the list of vehicle attribute categories based on the driver profile data, which includes the enriched driver profile data and the user-generated driver profile data; and
providing the enriched driver profile data, the user-generated driver profile data, the vehicle attribute data, and the driver attribute target values as input data to a trained neural network to generate at least one vehicle recommendation, wherein the trained neural network functions as a hybrid recommender system performing content-based filtering and collaborative filtering and wherein the neural network is trained on vehicle attribute data and driver profile data.

### Claim 15 (depends on 14)

The method of claim 14, further comprising determining at least one vehicle purchase recommendation based on the at least one vehicle recommendation, wherein the vehicle purchase recommendation includes at least one of a recommended purchase location, a recommended inventory, a recommended purchase time, and a recommended purchase method.