Machine learning based vehicle service recommendation system
A device receives current vehicle data that describes a current state of each of a plurality of components of the vehicle. The device accesses historical data describing prior services previously performed on the vehicle and applies the current vehicle data and the historical vehicle data to a trained machine learning model to obtain a set of recommended services, where the model was trained using labeled training data associated with additional vehicles having a threshold similarity to the vehicle. The machine learning model is trained to output a set of recommended services to be performed for a given vehicle based on inputs of given current vehicle data and given historical vehicle data for the given vehicle. The device outputs for display the set of recommended services to be performed on the vehicle to a user.
1. A method comprising:
receiving, by way of human input into a user interface relating to a vehicle inspection, current vehicle data describing a current state of each of a plurality of components of a vehicle;
accessing historical vehicle data associated with the vehicle, the historical vehicle data describing a set of services previously performed on the vehicle;
accessing user profile data indicating one or more vehicle servicing preferences of an owner of the vehicle;
applying the current vehicle data, the historical vehicle data, and the one or more vehicle servicing preferences of the owner of the vehicle as input to a supervised machine learning model configured to generate a set of recommended services to be performed on the vehicle, wherein the supervised machine learning model was trained by:
accessing labeled training data corresponding to a set of additional vehicles having at least a threshold similarity to the vehicle, wherein each label of each training example of the labeled training data indicates a set of services performed on a respective additional vehicle, and
training the supervised machine learning model using the labeled training data, the supervised machine learning model trained to output a set of recommended services to be performed for a given vehicle based on inputs of given current vehicle data, given historical vehicle data for the given vehicle, and given vehicle servicing preferences of a given owner of the given vehicle; and
outputting for display the set of recommended services to be performed on the vehicle to a user.
2. The method of claim 1 , further comprising:
ranking the set of recommended services to be performed on the vehicle;
providing for display selectable options, each selectable option associated with a recommended service in the set of recommended services;
determining that the user has interacted with a selectable option corresponding to a recommended service; and
responsive to determining that the user has interacted with the selectable option:
reranking a subset of the selectable options, and
updating the ranking based on the reranking, the updated ranking output for display along with a selectable option for each of a subset of the selectable options.
3. The method of claim 1 , further comprising:
generating for display to the user one or more interface elements, each of the one or more interface elements to allow a user to select one or more values associated with vehicle services;
in response to receiving an indication that a user interacted with at least one interface element of the one or more interface elements, optimizing the set of recommended services to be performed on the vehicle; and
updating the interface elements based on the optimized set of recommended services.
4. The method of claim 1 , further comprising:
filtering the set of recommended services based on one or more of: a value entered by an owner of the vehicle, a severity of service of each of the recommended services in the set of recommended services, and a likelihood the owner of the vehicle will accept a recommended service, wherein providing the set of recommended services to be performed on the vehicle comprises providing the filtered set of recommended services for display.
5. The method of claim 1 , further comprising:
determining a set of services a service provider has available to provide to vehicle owners; and
filtering the set of recommended services based on a set of services a service provider offers to vehicle owners.
6. The method of claim 1 , wherein the labeled training data includes one or more of: historical inspection information of the additional vehicles, inspection values of the additional vehicles, services performed on the additional vehicles, and services declined by customers of the additional vehicles.
7. The method of claim 1 , further comprising:
classifying the historical vehicle data into a plurality of predetermined service classifications, wherein the set of recommended services includes services corresponding to the predetermined service classifications.
8. The method of claim 1 , wherein the current vehicle data includes inspection information provided by a technician of the vehicle.
9. The method of claim 1 , further comprising: retraining the supervised machine learning model based on services of the set of recommended services that were accepted and declined by the user.
10. The method of claim 1 , wherein receiving the current vehicle data comprises receiving data from one or more sensors of the vehicle.
11. The method of claim 10 , wherein receiving the current vehicle data comprises receiving data from a computer of the vehicle, the computer operably coupled to the one or more sensors of the vehicle.
12. The method of claim 1 , wherein receiving the current vehicle data comprises receiving data from a client device of a service technician.
13. A non-transitory computer-readable medium comprising memory with instructions encoded thereon that, when executed, cause one or more processors to perform operations, the instructions comprising instructions to:
receive, by way of human input into a user interface relating to a vehicle inspection, current vehicle data describing a current state of each of a plurality of components of a vehicle;
access historical vehicle data associated with the vehicle, the historical vehicle data describing a set of services previously performed on the vehicle;
access user profile data indicating one or more vehicle servicing preferences of an owner of the vehicle;
apply the current vehicle data, the historical vehicle data, and the one or more vehicle servicing preferences of the owner of the vehicle as input to a supervised machine learning model configured to generate a set of recommended services to be performed on the vehicle, wherein the supervised machine learning model was trained by:
accessing labeled training data corresponding to a set of additional vehicles having at least a threshold similarity to the vehicle, wherein each label of each training example of the labeled training data indicates a set of services performed on a respective additional vehicle, and
training the supervised machine learning model using the labeled training data, the supervised machine learning model trained to output a set of recommended services to be performed for a given vehicle based on inputs of given current vehicle data, given historical vehicle data for the given vehicle, and given vehicle servicing preferences of a given owner of the given vehicle; and
output for display the set of recommended services to be performed on the vehicle to a user.
14. The non-transitory computer-readable medium of claim 13 , the instructions further comprising instructions to:
rank the set of recommended services to be performed on the vehicle;
provide for display selectable options, each selectable option associated with a recommended service in the set of recommended services;
determine that the user has interacted with a selectable option corresponding to a recommended service; and
responsive to determining that the user has interacted with the selectable option:
re-rank a subset of the selectable options, and
update the ranking based on the re-ranking, the updated ranking output for display along with a selectable option for each of a subset of the selectable options.
15. The non-transitory computer-readable medium of claim 13 , the instructions further comprising instructions to:
generate for display to the user one or more interface elements, each of the one or more interface elements to allow a user to select one or more values associated with vehicle services;
in response to receiving an indication that a user interacted with at least one interface element of the one or more interface elements, optimize the set of recommended services to be performed on the vehicle; and
update the interface elements based on the optimized set of recommended services.
16. The non-transitory computer-readable medium of claim 13 , the instructions further comprising instructions to:
filter the set of recommended services based on one or more of: a value entered by an owner of the vehicle, a severity of service of each of the recommended services in the set of recommended services, and a likelihood the owner of the vehicle will accept a recommended service, wherein providing the set of recommended services to be performed on the vehicle comprises providing the filtered set of recommended services for display.
17. The non-transitory computer-readable medium of claim 13 , the instructions further comprising instructions to:
determine a set of services a service provider has available to provide to vehicle owners; and
filter the set of recommended services based on a set of services a service provider offers to vehicle owners.
18. The non-transitory computer-readable medium of claim 13 , wherein the labeled training data includes one or more of: historical inspection information of the additional vehicles, inspection values of the additional vehicles, services performed on the additional vehicles, and services declined by customers of the additional vehicles.
19. The non-transitory computer-readable medium of claim 13 , the instructions further comprising instructions to:
classify the historical vehicle data into a plurality of predetermined service classifications, wherein the set of recommended services includes services corresponding to the predetermined service classifications.
20. A system comprising:
memory with instructions encoded thereon; and
one or more processors that, when executing the instructions, are caused to perform operations comprising:
receiving, by way of human input into a user interface relating to a vehicle inspection, current vehicle data describing a current state of each of a plurality of components of a vehicle;
accessing historical vehicle data associated with the vehicle, the historical vehicle data describing a set of services previously performed on the vehicle;
accessing user profile data indicating one or more vehicle servicing preferences of an owner of the vehicle;
applying the current vehicle data, the historical vehicle data, and the one or more vehicle servicing preferences of the owner of the vehicle as input to a supervised machine learning model configured to generate a set of recommended services to be performed on the vehicle, wherein the supervised machine learning model was trained by:
accessing labeled training data corresponding to a set of additional vehicles having at least a threshold similarity to the vehicle, wherein each label of each training example of the labeled training data indicates a set of services performed on a respective additional vehicle, and
training the supervised machine learning model using the labeled training data, the supervised machine learning model trained to output a set of recommended services to be performed for a given vehicle based on inputs of given current vehicle data, given historical vehicle data for the given vehicle, and given vehicle servicing preferences of a given owner of the given vehicle; and
outputting for display the set of recommended services to be performed on the vehicle to a user.