US 12,271,867 B1
Predicting resource lifecycles and managing resources in enterprise networks
Rebecca Anne Olmstead, Bloomington, IL (US); Amy Marie Stuepfert, Le Roy, IL (US); Julie Christine Wirtjes, Bloomington, IL (US); Cheryl L. Legan, Normal, IL (US); Bruce Lininger, Normal, IL (US); Gurnek Singh Mokha, Urbana, IL (US); Mathew Farley, Champaign, IL (US); Patrick Kan, Cupertino, CA (US); Sudesh Sahu, Normal, IL (US); and Socrates Krishnamurthy, Aurora, IL (US)
Assigned to State Farm Mutual Automobile Insurance Company, Bloomington, IL (US)
Filed by State Farm Mutual Automobile Insurance Company, Bloomington, IL (US)
Filed on Feb. 9, 2021, as Appl. No. 17/171,666.
Claims priority of provisional application 62/972,561, filed on Feb. 10, 2020.
Int. Cl. G06Q 10/20 (2023.01); G06Q 10/0631 (2023.01); G06Q 10/10 (2023.01)
CPC G06Q 10/20 (2013.01) [G06Q 10/06312 (2013.01); G06Q 10/10 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
generating a trained machine-learned prediction model by training a machine-learned prediction model using a training data set and a machine-learning algorithm, wherein training the machine-learned prediction model includes configuring the machine-learned prediction model, using the machine-learning algorithm, to identify relationships between resource lifecycles, resource vendors, and resource technological types;
identifying that a resource is connected to an enterprise network, the resource comprising at least one of hardware or software, being associated with a vendor, and being among a class of a plurality of classes, wherein:
individual classes in the plurality of classes are assigned by the trained machine-learned prediction model to a respective distinct technology type of a plurality of technology types; and
multiple resources sharing a technological type of the plurality of technology types are assigned to the class;
executing the trained machine-learned prediction model, using documentation associated with the vendor and the technological type as a first input, to generate an initial lifecycle of the resource as a first output, the initial lifecycle including an initial expected replacement time, wherein executing the trained machine-learned prediction model comprises:
the trained machine-learned prediction model performing natural language processing of the documentation to determine the vendor as model input, and
the trained machine-learned prediction model determining:
a mathematical significance associated with the vendor and the technological type, and
the initial lifecycle including the initial expected replacement time based on the mathematical significance;
storing the initial lifecycle in a database together with an identification of the resource;
receiving, from a resource tracking system, a data packet comprising an indication of at least one of a change in latency associated with the resource or a change in processor utilization associated with the resource;
determining that a trigger event associated with the resource has occurred based at least in part on the data packet;
executing the trained machine-learned prediction model, using the trigger event and the class as a second input, to:
generate a plurality of rules associated with the class, and
generate, based at least in part on the trigger event and at least one rule of the plurality of rules associated with the trigger event, an updated lifecycle of the resource, the updated lifecycle including an updated expected replacement time as a second output;
generating, based at least in part on the second output, a report indicating the updated lifecycle of the resource;
transmitting the report to a user device;
automatically replacing the resource in advance of the updated expected replacement time based at least in part on the report; updating the training data set with training data associating the vendor, the technological type, and the updated lifecycle of the resource; and
retraining the machine-learned prediction model using the updated training data set and the machine-learning algorithm.