US 10,389,739 B2Granted

System, method, and computer program for detecting regular and irregular events associated with various entities

AssigneeAmdocs Development Limited

InventorGad N. Solotorevsky

Life of this patent

Estimating the term…

Loading patent…

Abstract

A system, method, and computer program product are provided for detecting regular and irregular events associated with various entities. In operation, an event detection system detects at least one event associated with at least one entity. The event detection system identifies the at least one entity. Additionally, the event detection system identifies one or more historical patterns associated with the at least one event. Further, the event detection system determines whether the at least one event is anomalous based on the one or more historical patterns and details associated with the at least one event. Moreover, the event detection system performs at least one action based on whether the at least one event is determined to be anomalous.

Claims16

Dependent claims

A method, comprising:

learning, by an event detection system, patterns of events across a plurality of entities, utilizing historical events identified for the plurality of entities;
storing, by the event detection system, the patterns of events including storing for each pattern of events:
key features of the pattern of events, and
an indicator of one or more entities associated with the pattern of events;
building, by the event detection system, a scoring model for the patterns of events, the scoring model indicating scores for variances to the patterns of events;
detecting, by the event detection system, a sequence of events including an event to authenticate;
identifying, by the event detection system, one or more of the stored patterns associated with the event to authenticate;
using the scoring model to score, by the event detection system, the event to authenticate based on the detected sequence of events and the one or more of the stored patterns associated with the event to authenticate;
determining, by the event detection system, whether the event to authenticate is anomalous based on the score of the event;
and
performing, by the event detection system, at least one first action responsive to determining that the event is anomalous;
and
performing, by the event detection system, at least one second action responsive to determining that the event is not anomalous.

10 more elements

Dependencies are inferred from the claim text.

Facts

Application: 15/482,368Filed 7 Apr 2017
Published: US 2018/0295141 A111 Oct 2018
Granted: 20 Aug 2019
Term ends: Not estimatedThe source records don’t support an estimate.

Classification11 codes

H04L 63/1416
By monitoring network traffic: event detection, e.g. attack signature detection
G06F 21/554
Detecting local intrusion or implementing counter-measures: involving event detection and direct action
G06N 5/04
Computing arrangements using knowledge-based models: inference or reasoning models
G06N 20/00
Machine learning