US 11,929,985 B2
Network-based authentication rule cleaning and optimization
Chaoyun Chen, San Francisco, CA (US); Quan Li, Fremont, CA (US); and Zifang Huang, Redwood City, CA (US)
Assigned to The Western Union Company, Denver, CO (US)
Filed by The Western Union Company, Denver, CO (US)
Filed on Mar. 24, 2023, as Appl. No. 18/189,757.
Application 18/189,757 is a continuation of application No. 16/837,298, filed on Apr. 1, 2020, granted, now 11,627,113.
Prior Publication US 2023/0328036 A1, Oct. 12, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. H04L 9/40 (2022.01); G06N 7/01 (2023.01); G06N 20/00 (2019.01); H04L 41/0893 (2022.01)
CPC H04L 63/0263 (2013.01) [G06N 7/01 (2023.01); G06N 20/00 (2019.01); H04L 41/0893 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
one or more processing devices; and
memory coupled with the one or more processing devices and storing instructions that, when executed by the one or more processing devices, cause the one or more processing devices to perform operations comprising:
determining a plurality of rules that were previously used to evaluate a plurality of financial transactions;
generating a cleaning coefficient for a set of one or more rules of the plurality of rules, wherein the cleaning coefficient indicates how the set of one or more rules of the plurality of rules is performing, where the cleaning coefficient is based at least in part on whether the set of one or more rules prevents one or more financial transactions of the plurality of financial transactions;
identifying the set of one or more rules from the plurality of rules to eliminate from the plurality of rules, based at least in part on the cleaning coefficient, wherein a machine learning module is used to identify the set of one or more rules or is trained using the set of one or more rules;
eliminating at least one rule of the set of one or more rules from the plurality of rules to generate a modified set of rules, based on the identified set of one or more rules to eliminate;
authenticating or preventing a transaction based on the modified set of rules;
identifying a set of one or more attributes associated with the identified set of one or more rules to eliminate; and
training the machine learning module based at least in part on the identified set of one or more attributes associated with the identified set of one or more rules to eliminate.