US 12,243,645 B2
Iterated training of machine models with deduplication
Pritesh J. Shah, Paramus, NJ (US); Christopher R. Markson, Hawthorne, NJ (US); and Logan R. Meltabarger, St. Charles, MO (US)
Assigned to Express Scripts Strategic Development, Inc., St. Louis, MO (US)
Filed by Express Scripts Strategic Development, Inc., St. Louis, MO (US)
Filed on Dec. 18, 2023, as Appl. No. 18/543,804.
Application 18/543,804 is a continuation of application No. 17/363,605, filed on Jun. 30, 2021, granted, now 11,848,101.
Application 17/363,605 is a continuation of application No. 15/655,647, filed on Jul. 20, 2017, granted, now 11,087,880, issued on Aug. 10, 2021.
Application 15/655,647 is a continuation in part of application No. 15/498,048, filed on Apr. 26, 2017, abandoned.
Claims priority of provisional application 62/327,994, filed on Apr. 26, 2016.
Prior Publication US 2024/0120103 A1, Apr. 11, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G16H 50/20 (2018.01); G06N 20/00 (2019.01); G06Q 50/22 (2024.01); G16H 10/60 (2018.01)
CPC G16H 50/20 (2018.01) [G06N 20/00 (2019.01); G06Q 50/22 (2013.01); G16H 10/60 (2018.01)] 18 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
defining model attributes including a training iteration value that defines a set of training iterations to be used in machine learning to associate portions of feedback data with a set of topic groups based on similarities in concepts conveyed in the feedback data, wherein
the feedback data includes confidential information;
removing at least some of the confidential information from the feedback data, wherein the removed confidential information includes at least one of: pharmacy claims data, drug code numbers of medications, financial costs of the medications, copay amounts associated with a set of prescription benefit plans, or member eligibilities associated with the set of prescription benefit plans;
receiving a topic model number selection that indicates a subset of the set of topic groups;
using machine learning to train a machine model based on the model attributes and the topic model number selection, wherein:
the machine model is trained by defining relationships between selected concepts for each training iteration of the set of training iterations, and
for each iteration of the set of training iterations, the feedback data is transformed by (i) applying an ontology to reduce terms in the feedback data by eliminating duplicative entries in the feedback data and (ii) sorting the portions of the feedback data having a relationship with each other into a topic group of the set of topic groups, wherein the set of training iterations dictates how the set of topic groups is defined and associated with the portions of the feedback data; and
creating at least one software application using the trained machine model, wherein the at least one software application is configured to generate a display based on the trained machine model, wherein the display visually depicts information of the set of topic groups.