US 12,149,401 B2
Identifying persistent anomalies for failure prediction
Seema Nagar, Bangalore (IN); Pooja Aggarwal, Bengaluru (IN); Dipanwita Guhathakurta, Kolkata (IN); Rohan R Arora, Danbury, CT (US); Amitkumar Manoharrao Paradkar, Mohegan Lake, NY (US); Larisa Shwartz, Greenwich, CT (US); Bing Zhou, Rye, NY (US); and Noah Zheutlin, White Plains, NY (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Nov. 23, 2021, as Appl. No. 17/456,168.
Prior Publication US 2023/0164035 A1, May 25, 2023
Int. Cl. H04L 41/0631 (2022.01); G06N 3/02 (2006.01); H04L 41/14 (2022.01); H04L 41/147 (2022.01); H04L 41/16 (2022.01); H04L 43/06 (2022.01)
CPC H04L 41/064 (2013.01) [G06N 3/02 (2013.01); H04L 41/145 (2013.01); H04L 41/147 (2013.01); H04L 41/16 (2013.01); H04L 43/06 (2013.01)] 24 Claims
OG exemplary drawing
 
1. A computer-implemented method for identifying persistent anomalies for failure prediction, the method comprising:
receiving a time series data stream;
receiving a predetermined number N and a predetermined number M which is a fraction of N, wherein N and M are positive integers;
segmenting the time series data stream into N consecutive sliding windows;
performing supervised persistent anomaly detection to determine whether anomalies across the N consecutive sliding windows are persistent, based on a prediction that labels that arm indicative of an outage occurrence are in at least M sliding windows, by using a binary classification model;
performing unsupervised persistent anomaly detection to determine whether the anomalies across the N consecutive sliding windows are persistent, based on clustering the anomalies and growth of clusters of the anomalies in at least M sliding windows; and
combining results of the supervised persistent anomaly detection and results of the unsupervised persistent anomaly detection to determine persistent anomalies.