Hierarchical pattern matching for anomaly detection in time series

As companies rely on an ever increasing number of connected devices for their day to operations, a need arises automated anomaly detectors constantly observe crucial device metrics in real time prevent downtime and data loss. production environments tend monitor huge amount these metrics, it prevents current state-of-the-art techniques be deployed as the required computational resources is too high. This paper proposes lightweight detection method that can without reduction accuracy. The approach works fully online, does not require extensive history set kept memory. benchmarked publicly available Numenta dataset, well network monitoring dataset from different provided by management solution vendor. These benchmarks show proposed technique very competitive with exceeding applicability.