US 12,086,886 B2
Machine learning for insurance applications
Anthony DiFranco, Twinsburg, OH (US)
Assigned to AMTRUST FINANCIAL SERVICES, INC., New York, NY (US)
Filed by AmTrust Financial Services, Inc., New York, NY (US)
Filed on Nov. 23, 2021, as Appl. No. 17/533,621.
Claims priority of provisional application 63/117,747, filed on Nov. 24, 2020.
Prior Publication US 2022/0164893 A1, May 26, 2022
Int. Cl. G06Q 40/08 (2012.01); G06F 40/40 (2020.01); G06N 20/00 (2019.01); G06Q 30/0204 (2023.01)
CPC G06Q 40/08 (2013.01) [G06F 40/40 (2020.01); G06N 20/00 (2019.01); G06Q 30/0205 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A method, comprising:
receiving, over a computer network, data corresponding to a first insurance application by an insurance application portal of a server from a user device, wherein the data corresponding to the first insurance application includes data corresponding to a first insurance line of business of an insurance carrier, and wherein the data corresponding to the first insurance application further includes a first set of customer information;
determining, by the server, a second insurance line of business having a second insurance application in accordance with at least one of the first set of customer information or the first insurance application;
analyzing, by the server, the second insurance application to determine a second set of customer information corresponding to the second insurance application;
determining, by the server, the first set of customer information is incomplete for the second insurance application for the determined second insurance line of business;
building, by the server, a prediction model for inferring customer information, the building comprising:
determining a set of customers of the insurance carrier having substantially similar customer information to the first set of customer information via at least one of a matching algorithm or a similarity scoring and threshold technique, wherein the substantially similar customer information includes previous requests of the second line of business from the set of customers,
building a training dataset for the prediction model in accordance with the determined set of customers, and
training the prediction model using supervised machine learning in accordance with the training dataset to estimate values of a data field or a data attribute of customer information;
inferring, by the server, a value of at least one of the data field or the data attribute corresponding to missing data of the second set of customer information using the prediction model;
completing, by the server, the second set of customer information in accordance with the inferred value of the at least one of the data field or data attribute;
completing, by the server, the second insurance application in accordance with the completed second set of customer information including the inferred value of the at least one of the data field or data attribute; and
communicating, by the server, the completed second insurance application to a user interface of the user device via the computer network.