US 11,704,583 B2
Machine learning and validation of account names, addresses, and/or identifiers
Donald J. McQueen, Leesburg, VA (US); and Lachlan A. Maxwell, Ashburn, VA (US)
Assigned to Yahoo Assets LLC, New York, NY (US)
Filed by Yahoo Assets LLC, Dulles, VA (US)
Filed on Aug. 14, 2020, as Appl. No. 16/994,221.
Application 16/994,221 is a continuation of application No. 15/895,520, filed on Feb. 13, 2018, granted, now 10,789,537.
Application 15/895,520 is a continuation of application No. 14/282,097, filed on May 20, 2014, granted, now 9,928,465, issued on Mar. 27, 2018.
Prior Publication US 2020/0380395 A1, Dec. 3, 2020
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 5/048 (2023.01); G06N 20/00 (2019.01); H04L 9/40 (2022.01); H04L 51/212 (2022.01); G06N 7/01 (2023.01)
CPC G06N 5/048 (2013.01) [G06N 7/01 (2023.01); G06N 20/00 (2019.01); H04L 51/212 (2022.05); H04L 63/126 (2013.01); H04L 63/1466 (2013.01); H04L 63/0236 (2013.01); H04L 63/0245 (2013.01); H04L 63/1425 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method for determining if an electronic account identifier is computer-generated, comprising:
receiving the electronic account identifier;
analyzing the electronic account identifier to determine a plurality of fragments comprising hashed or truncated identifier fragments of a predetermined character length range;
comparing the plurality of fragments to determine one or more alphanumeric features of at least one fragment;
comparing the at least one fragment with a second plurality of fragments associated with a plurality of electronic account identifiers to determine a percentile of commonness of the at least one fragment;
determining, by a computer, if the electronic account identifier is computer-generated beyond a predetermined confidence threshold based on the determined one or more features of the at least one fragment;
transmitting the determined computer generated electronic account identifier to a probabilistic classifier model, wherein the probabilistic classifier model is a trained machine learning system; and
training the probabilistic classifier model further based on the transmitted electronic account identifier.