	US 12,450,503 B2
	Automatically-generated labels for time series data and numerical lists to use in analytic and machine learning systems
Amr Fawzy Fahmy, Foxboro, MA (US); Sreeji Krishnan Das, Fremont, CA (US); Adrienne Wong, Redwood City, CA (US); Jae Young Yoon, San Mateo, CA (US); Dhileeban Kumaresan, Foster City, CA (US); and Eric L. Sutton, Naperville, IL (US)
Assigned to Oracle International Corporation, Redwood Shores, CA (US)
Filed by Oracle International Corporation, Redwood Shores, CA (US)
Filed on Dec. 15, 2023, as Appl. No. 18/542,282.
Application 18/542,282 is a continuation of application No. 16/856,790, filed on Apr. 23, 2020, granted, now 11,887,015.
Claims priority of provisional application 62/900,418, filed on Sep. 13, 2019.
Prior Publication US 2024/0119319 A1, Apr. 11, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 5/04 (2023.01); G06F 16/24 (2019.01); G06N 20/00 (2019.01)

CPC G06N 5/04 (2013.01) [G06F 16/24 (2019.01); G06N 20/00 (2019.01)]

20 Claims

1. A method comprising:

receiving a query that includes a composite label associated with an event, wherein the composite label is associated with a particular pattern of different labels that include at least a first label and a second label;

responsive to receiving the query, executing at least one of a vertical or horizontal search for the particular pattern of different labels in a set of labels automatically assigned to data points in at least one time series;

based on at least one of the vertical or horizontal search, identifying the particular pattern of labels associated with the composite label, wherein the particular pattern is identified based at least in part on at least one occurrence of a first label assigned to a first data point and a second label assigned to a second data point within a threshold number of labels from each other in the at least one time series; and

responsive to identifying the particular pattern of labels associated with the composite label, triggering a responsive action.