US 11,876,825 B1
System, method, and computer program for using metric valuations in anomaly scoring
Gilli Shama, Ra'anana (IL); and Nadeem Azaizah, Dabburiya (IL)
Assigned to AMDOCS DEVELOPMENT LIMITED, Limassol (CY)
Filed by Amdocs Development Limited, Limassol (CY)
Filed on Jul. 29, 2021, as Appl. No. 17/389,175.
Int. Cl. H04L 9/40 (2022.01); H04L 41/16 (2022.01); H04L 41/147 (2022.01); H04L 41/142 (2022.01)
CPC H04L 63/1433 (2013.01) [H04L 41/142 (2013.01); H04L 41/147 (2013.01); H04L 41/16 (2013.01); H04L 63/1425 (2013.01)] 8 Claims
OG exemplary drawing
 
1. A non-transitory computer-readable media storing computer instructions which when executed by one or more processors of a device cause the device to:
predict a state expected at a future time t, using a machine learning model, wherein the predicted state includes a series of scores predicted for a corresponding series of metrics at the future time t, wherein the metrics each correspond to a different parameter of an application or system component being monitored for the purpose of detecting anomalies in a functionality or configuration of the application or system component;
determine a weighted score for the predicted state by:
for each score in the series of scores, multiplying the score to a weight defined for a metric in the series of metrics that corresponds to the score to form a weighted score for the metric, wherein the weight defined for the metric varies based on time and is accordingly dependent on the future time t, and
determining the weighted score for the predicted state as an average of the weighted scores formed for the metrics;
cause at least one action to be performed, based on the weighted score for the predicted state, including:
determining that the weighted score for the predicted state is outside of a defined range indicative of a desired anomaly and thus that the weighted score for the predicted state indicates an undesired anomaly, and
causing the at least one action to be performed responsive to determining that the weighted score for the predicted state indicates the undesired anomaly.