US 12,249,423 B2
Machine learning system for generating predictions according to varied attributes
Armand E. Prieditis, McKinleyville, CA (US); and John E. Paton, Raleigh, NC (US)
Assigned to Cigna Intellectual Property, Inc., Wilmington, DE (US)
Filed by Cigna Intellectual Property, Inc., Wilmington, DE (US)
Filed on Dec. 18, 2023, as Appl. No. 18/543,670.
Application 18/543,670 is a continuation of application No. 18/073,698, filed on Dec. 2, 2022, granted, now 11,848,104.
Application 18/073,698 is a continuation of application No. 16/731,316, filed on Dec. 31, 2019, granted, now 11,521,744, issued on Dec. 6, 2022.
Claims priority of provisional application 62/862,845, filed on Jun. 18, 2019.
Prior Publication US 2024/0120102 A1, Apr. 11, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G16H 50/20 (2018.01); G06F 17/16 (2006.01); G06F 17/18 (2006.01); G06N 20/00 (2019.01); G16H 50/30 (2018.01)
CPC G16H 50/20 (2018.01) [G06F 17/16 (2013.01); G06F 17/18 (2013.01); G06N 20/00 (2019.01); G16H 50/30 (2018.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
receiving a group of sets selected from a plurality of stored sets, wherein each set in the group of sets has values for immutable attributes that match a set of values for at least one mutable attribute in a prediction request;
generating a deviation model based on the group of sets;
generating, using the deviation model, a plurality of sets of mutable attribute values;
automatically selecting a neural network from a data store based on at least one of the plurality of sets of mutable attribute values, wherein:
the neural network includes a set of layers,
each layer of the set of layers includes a set of nodes,
a first layer of the set of layers is an input layer configured to receive inputs at the set of nodes of the input layer, and
a last layer of the set of layers is an output layer configured to output, from the neural network, outputs modified by a preceding layer of the set of layers; and
generating a likelihood corresponding to the at least one mutable attribute in the prediction request using the neural network.