US 11,675,799 B2
Anomaly detection system
Francesco Pierri, Salerno (IT); Ioana Giurgiu, Zurich (CH); Monney Serge, Pully (CH); and Mitch Gusat, Langnau a.A. (CH)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on May 5, 2020, as Appl. No. 16/866,884.
Prior Publication US 2021/0349897 A1, Nov. 11, 2021
Int. Cl. G06F 11/34 (2006.01); G06F 16/2458 (2019.01); G06N 5/04 (2023.01); G06N 20/00 (2019.01); G06F 11/07 (2006.01)
CPC G06F 16/2474 (2019.01) [G06F 11/079 (2013.01); G06F 11/3447 (2013.01); G06F 11/3452 (2013.01); G06N 5/04 (2013.01); G06N 20/00 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A method for anomaly detection, the method comprising:
receiving, by one or more processors, a first set of time series from a data source, wherein the data source provides sensor data and a timestamping of the sensor data as the first set of time series, the sensor data comprising values of a first group of measurands;
identifying, by one or more processors, unexpected values of monitoring measurands in a monitored time series utilizing an anomaly detection algorithm;
determining, by one or more processors, that values of a second group of one or more of the measurands of a subset of the received sensor data indicates an anomaly by executing the anomaly detection algorithm on the received time series;
sending, by one or more processors, anomalous data indicative of the subset of sensor data to a root cause analysis system;
receiving, by one or more processors, a root cause analysis feedback from the root cause analysis system, the root cause analysis feedback being indicative of a result of a root cause analysis of the subset of sensor data, the root cause analysis feedback comprising a third group of the measurands; and
training, by one or more processors, the anomaly detection algorithm based on a difference between the third groups of measurands and the second groups of measurands to predict values of monitoring measurands deviating from a normal behavior.