US 12,314,125 B2
Systems and methods for automated anomaly detection in univariate time-series
Manjunath Channappagoudar, Bangalore (IN); Arvind Shyam Verma, Bengaluru (IN); Mohit Choudhary, Bentonville, AR (US); Jingying Zhang, Beaumont, TX (US); Juan Gomez, Bentonville, AR (US); and Lokesh Kumar Sambasivan, Tirupati (IN)
Assigned to Walmart Apollo, LLC, Bentonville, AR (US)
Filed by Walmart Apollo, LLC, Bentonville, AR (US)
Filed on Jul. 27, 2023, as Appl. No. 18/360,704.
Prior Publication US 2025/0036508 A1, Jan. 30, 2025
Int. Cl. G06F 11/00 (2006.01); G06F 11/07 (2006.01); G06F 16/2458 (2019.01)
CPC G06F 11/079 (2013.01) [G06F 16/2474 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A system, comprising:
a processor; and
a non-transitory memory storing instructions that, when executed, cause the processor to:
receive a time-series dataset;
implement an iterative process to define a set of segments from the time-series dataset that:
identifies a set of changepoints in the time-series dataset based on a changepoint type and a sensitivity parameter;
determines whether the set of segments defined by the set of changepoints satisfy at least one predetermined threshold criteria;
when the set of segments do not satisfy the at least one predetermined threshold criteria, modifies the sensitivity parameter; and
when the set of segments satisfy the at least one predetermined threshold criteria, outputs the set of segments;
determine a segment-specific anomaly detection threshold for each segment in the set of segments;
generate a set of anomaly-flagged segments, wherein the set of anomaly-flagged segments are generated by an anomaly detection process based on the segment-specific anomaly detection threshold for a corresponding segment; and
combine the set of anomaly-flagged segments to generate an anomaly-flagged time-series dataset.