US 11,809,499 B2
Machine learning segmentation methods and systems
Marcio Oliveira Almeida, Stittsville (CA); Seyednaser Nourashrafeddin, Ottawa (CA); Jean-François Dubeau, Ottawa (CA); Ivy Blackmore, Ottawa (CA); and Zhen Lin, Kanata (CA)
Assigned to Kinaxis Inc., Ottawa (CA)
Filed by Kinaxis Inc., Ottawa (CA)
Filed on Apr. 14, 2020, as Appl. No. 16/848,266.
Application 16/848,266 is a continuation in part of application No. 16/837,182, filed on Apr. 1, 2020, granted, now 11,537,825.
Application 16/837,182 is a continuation in part of application No. 16/697,620, filed on Nov. 27, 2019.
Application 16/837,182 is a continuation in part of application No. 16/599,143, filed on Oct. 11, 2019, granted, now 11,526,899.
Claims priority of provisional application 62/915,076, filed on Oct. 15, 2019.
Prior Publication US 2021/0109969 A1, Apr. 15, 2021
Int. Cl. G06Q 10/06 (2023.01); G06Q 30/02 (2023.01); G06Q 30/06 (2023.01); G06F 16/906 (2019.01); G06N 20/00 (2019.01); G06F 18/25 (2023.01)
CPC G06F 16/906 (2019.01) [G06F 18/251 (2023.01); G06N 20/00 (2019.01)] 21 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
receiving, by a segmentation engine, data associated with each item of a plurality of items, each item associated with one or more attributes;
engineering, by the segmentation engine, features associated with one or more signals, the one or more signals comprising either: i) one or more internal signals associated with the data; or ii) one or more internal signals associated with the data and one or more external signals;
selecting, by the segmentation engine, a set of the engineered features;
training, by the segmentation engine, a plurality of cluster-based machine learning models on the set of the engineered features;
generating, by the segmentation engine, a plurality of segmentations of the plurality of items;
selecting, by the segmentation engine, a first segmentation of the plurality of segmentations based on one or more cluster-based metrics;
visualizing, by a visual user interface, the first segmentation of the plurality of segmentations; and
amending, in response to a user input received via the visual interface, one or more segments within the first segmentation of the plurality of segmentations;
wherein amending comprises:
retraining the plurality of cluster-based machine learning models; and
selecting, by the segmentation engine, a second segmentation of the plurality of amended segmentations based on one or more cluster-based metrics.