Maintenance targeting generator and model feedback computing system and related methods
AssigneeState Farm Mutual Automobile Insurance Company
InventorSean Hickey
Life of this patent
Estimating the term…
Abstract
A computer system is provided. The computer system may be configured to generate maintenance targetings and model feedback based at least in part upon sample maintenance data and user requests including subject building components and inquiries related thereto.
Claims20
A maintenance model feedback computing device for targeted component repair or maintenance instruction comprising at least one memory device and at least one processor, the at least one processor programmed to:
- ingest sample maintenance data for a plurality of sample homes, the sample maintenance data including maintenance histories corresponding to each of the plurality of sample homes and building components thereof;
- train a machine-learning model using the sample maintenance data as initial training data set;
- receive, from a user computing device, a request message including an inquiry related to a subject building component;
- execute the trained machine-learning model based upon one or more data elements in the inquiry to receive a model output of at least one likely maintenance target responsive to the inquiry;
- present the at least one likely maintenance target responsive to the inquiry;
- receive feedback, from the user computing device, associated with an efficacy of the at least one likely maintenance target responsive to the inquiry;
- and
- re-train the trained machine-learning model using the received feedback.
A computer-implemented method for maintenance model feedback including targeted component repair or maintenance instruction, the method implemented using a maintenance model feedback computing device including at least one memory device and at least one processor, the method comprising:
- ingesting sample maintenance data for a plurality of sample homes, the sample maintenance data including maintenance histories corresponding to each of the plurality of sample homes and building components thereof;
- training a machine-learning model using the sample maintenance data as initial training data set;
- receiving, from a user computing device, a request message including an inquiry related to a subject building component;
- executing the trained machine-learning model based upon one or more data elements in the inquiry to receive a model output of at least one likely maintenance target responsive to the inquiry;
- presenting the at least one likely maintenance target responsive to the inquiry;
- receiving feedback, from the user computing device, associated with an efficacy of the at least one likely maintenance target responsive to the inquiry;
- and
- re-training the trained machine-learning model using the received feedback.
Dependencies are inferred from the claim text.