Overview / Time series anomaly discovery
Time series anomaly discovery
Variable-length discord (surprise pattern) discovery with the Rare Rule Anomaly algorithm and the rule density curve.
This module discusses variable-length anomaly (i.e., discord, surprise pattern) discovery in time series using GrammarViz 3.0. Background on the SAX building blocks — z-normalization, PAA, and the discretization itself — lives on the SAX-VSM sister site.
tutorial
Time series anomaly detection using GUI
Discover anomalous time series patterns using the GrammarViz 3.0 graphical user interface
tutorialTime series anomaly detection using CLI
Discover anomalous time series patterns using the GrammarViz 3.0 command line interface
readingTemporal anomaly discovery
A survey of previous work on time series anomaly discovery, from discords and HOT SAX to compression-based approaches.
Further reading
link
Entropy journal — Special Issue on Kolmogorov Complexity