Maximally Divergent Intervals for Anomaly Detection

10/21/2016

∙

We present new methods for batch anomaly detection in multivariate time series. Our methods are based on maximizing the Kullback-Leibler divergence between the data distribution within and outside an interval of the time series. An empirical analysis shows the benefits of our algorithms compared to methods that treat each time step independently from each other without optimizing with respect to all possible intervals.

READ FULL TEXT

Maximally Divergent Intervals for Anomaly Detection

Greenhouse: A Zero-Positive Machine Learning System for Time-Series Anomaly Detection

Multivariate Time-series Anomaly Detection via Graph Attention Network

CRATOS: Cogination of Reliable Algorithm for Time-series Optimal Solution

Detecting Regions of Maximal Divergence for Spatio-Temporal Anomaly Detection

Anomaly Detection in Predictive Maintenance: A New Evaluation Framework for Temporal Unsupervised Anomaly Detection Algorithms

Anomaly Attribution of Multivariate Time Series using Counterfactual Reasoning

tegdet: An extensible Python Library for Anomaly Detection using Time-Evolving Graphs

Maximally Divergent Intervals for Anomaly Detection

Related Research

Greenhouse: A Zero-Positive Machine Learning System for Time-Series Anomaly Detection

Multivariate Time-series Anomaly Detection via Graph Attention Network

CRATOS: Cogination of Reliable Algorithm for Time-series Optimal Solution

Detecting Regions of Maximal Divergence for Spatio-Temporal Anomaly Detection

Anomaly Detection in Predictive Maintenance: A New Evaluation Framework for Temporal Unsupervised Anomaly Detection Algorithms

Anomaly Attribution of Multivariate Time Series using Counterfactual Reasoning

tegdet: An extensible Python Library for Anomaly Detection using Time-Evolving Graphs