US 12,093,902 B2
Enhanced identification and management of cable network outages
Amrit Shaswat, Atlanta, GA (US); Brian Stublin, Atlanta, GA (US); Sarah Lau, Atlanta, GA (US); and Brad Demerich, Atlanta, GA (US)
Assigned to Cox Communications, Inc., Atlanta, GA (US)
Filed by Cox Communications, Inc., Atlanta, GA (US)
Filed on Jul. 29, 2021, as Appl. No. 17/388,375.
Prior Publication US 2023/0032163 A1, Feb. 2, 2023
Int. Cl. G06Q 10/20 (2023.01); G06N 20/20 (2019.01); G06Q 10/0639 (2023.01); H04L 67/10 (2022.01)
CPC G06Q 10/20 (2013.01) [G06N 20/20 (2019.01); G06Q 10/0639 (2013.01); H04L 67/10 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for managing cable network outages, the method comprising:
receiving, by a system, first data indicative of a first cable system outage;
instantiating, by the system, based on receiving the first data, a first serverless computing instance that:
determines that the first data deviate from cable outage criteria;
generates event data indicative of the first cable system outage based on determining that the first data deviate from the cable outage criteria;
compares the event data to second data indicative of at least one of a planned cable system outage or a second cable system outage in a same geographic area as the first cable system outage; and
generates, based on comparing the event data to the second data, a flag indicative of the at least one of the planned cable system outage or the second cable system outage;
instantiating, by the system, based on generation of the event data, a second serverless computing instance that:
calls a trained machine learning model that has been re-trained, using performance data indicative of whether previous event data corresponded to actionable outage events, to adjust thresholds or delay times used to determine whether the event data are indicative of actionable events;
inputs the event data and the flag to the re-trained machine learning model; and
generates, using the event data and the flag as inputs to the re-trained machine learning model using the adjusted thresholds or delay times, a score indicative of a probability that the first cable system outage is repairable by a technician;
determining, by the system, that the score is below a score threshold; and
determining, by the system, based on the score being below the score threshold, not to send the event data to a first system for assigning the technician to repair the first cable system outage.