Abstract: A method includes receiving network data describing operation of a network including a plurality of anomalies; clustering the network data to obtain clusters of groups of correlated anomalies; responsive to labeling the clusters of groups of correlated anomalies, utilizing the labels for the network data to train a model for automatic classification; and providing the model for automatic classification of additional network data. The clusters can be described through interpretable clustering (tree learning on cluster labels) (usage for description) or the cluster of interest can be deployed (again, tree learning on cluster labels) (usage for automatic classification).

Abstract: An example embodiment may involve obtaining training time series data spanning an observation time window and comprising a series of values of a metric at regularly-spaced sample points in time, and analyzing the training time series data to determine one of a periodicity or a pseudo-periodicity across a plurality of consecutive sub-windows, each equal in size to a reference time period and each spanned by the same number N of sample points of metric values. A reference pattern corresponding to a model time series having no anomalies, as well as a reference threshold, may be determined and stored. Runtime time series data may then be obtained and time aligned with the reference pattern. Deviations between the runtime time series and the reference pattern may be identified as anomalies if they exceed the reference threshold. Identified anomalies may be displayed in a display device.

Abstract: Systems and methods for detection, characterization, prediction, and next occurrence prediction of approximately periodic chain of events with missing occurrences using pattern recognition obtaining data from monitoring a system, wherein the data includes a plurality of records each includes at least a start time and a unique identifier; analyzing the plurality of records to detect a periodic chain of events, wherein the periodic chain of events includes clear or approximate periodicity that is detected based on a plurality of parameters including some missing occurrences therein; converting the periodic chain of events into a binary sequence with each bit representing a time bin and having a value based on a presence or absence of an event in the time bin; and analyzing the binary sequence to recognize a pattern therein to determine a next occurrence of an event in the periodic chain of events.

Abstract: The invention concerns a method and a system for determining root-cause diagnosis of events occurring during the operation of a communication network comprising monitoring time signals representative of the operation of the network to detect the occurrence of an event relative to the network traffic, and for each detected event, during the duration of said event obtaining distributions of data on several dimensions of the network linked to said event, automatically determining an event root-cause diagnosis of the detected event, called single event diagnosis, comprising at least one element of said distributions, an element being a value taken by a network dimension having a contribution in said distributions of data, the single event diagnosis determination using rules of business logic configuration organized hierarchically, which are applied according to said hierarchy to select at least one element of said distributions, the selection of more than one element comprising machine learning clustering.

Abstract: Systems and methods for detection, characterization, and prediction of real-time events having approximate periodicity include detection of events from raw data that are approximately periodic. The detection includes analyzing raw data to determine approximately periodic chains of events. The raw data can be related to network management systems, financial monitoring systems, medical monitoring, seismic activity monitoring, or any system that performs some management or monitoring of an underlying system or network having time lasting events. The detected approximately periodic events could be characterized and presented in natural language as well as used for prediction of future events via supervised machine learning.

Abstract: Visualization systems and methods include, responsive to obtaining data related to one or more events occurring during operation of a system, displaying information about a time-based component of the one or more events on a timeline; displaying fixed information about a timeless component of the one or more events on the timeline; and responsive to user input to modify a timescale of the timeline, adjusting display of the timeline by adjusting the information of the one or more events and maintaining the fixed information of the one or more events.

Abstract: Systems and methods for detection, characterization, and prediction of real-time events having approximate periodicity include detection of events from raw data that are approximately periodic. The detection includes analyzing raw data to determine approximately periodic chains of events. The raw data can be related to network management systems, financial monitoring systems, medical monitoring, seismic activity monitoring, or any system that performs some management or monitoring of an underlying system or network having time lasting events. The detected approximately periodic events could be characterized and presented in natural language as well as used for prediction of future events via supervised machine learning.

Abstract: The invention concerns a method and a system for determining root-cause diagnosis of events occurring during the operation of a communication network comprising monitoring time signals representative of the operation of the network to detect the occurrence of an event relative to the network traffic, and for each detected event, during the duration of said event obtaining distributions of data on several dimensions of the network linked to said event, automatically determining an event root-cause diagnosis of the detected event, called single event diagnosis, comprising at least one element of said distributions, an element being a value taken by a network dimension having a contribution in said distributions of data, the single event diagnosis determination using rules of business logic configuration organized hierarchically, which are applied according to said hierarchy to select at least one element of said distributions, the selection of more than one element comprising machine learning clustering.

Abstract: Visualization systems and methods include, responsive to obtaining data related to one or more events occurring during operation of a system, displaying information about a time-based component of the one or more events on a timeline; displaying fixed information about a timeless component of the one or more events on the timeline; and responsive to user input to modify a timescale of the timeline, adjusting display of the timeline by adjusting the information of the one or more events and maintaining the fixed information of the one or more events.

Abstract: An example embodiment may involve obtaining training time series data spanning an observation time window and comprising a series of values of a metric at regularly-spaced sample points in time, and analyzing the training time series data to determine one of a periodicity or a pseudo-periodicity across a plurality of consecutive sub-windows, each equal in size to a reference time period and each spanned by the same number N of sample points of metric values. A reference pattern corresponding to a model time series having no anomalies, as well as a reference threshold, may be determined and stored. Runtime time series data may then be obtained and time aligned with the reference pattern. Deviations between the runtime time series and the reference pattern may be identified as anomalies if they exceed the reference threshold. Identified anomalies may be displayed in a display device.