The term "pattern-based anomaly detection" is particularly at home in the fields of artificial intelligence, big data and smart data as well as cybercrime and cybersecurity. The aim is to automatically detect unusual or suspicious processes in large amounts of data.
Normally, many processes, such as network connections in the office or production steps in a factory, always follow similar patterns. Pattern-based anomaly detection uses artificial intelligence or software to learn this "normality" first. It then constantly analyses new data and sounds the alarm as soon as something unusual happens - for example, if a computer suddenly sends much larger amounts of data abroad than usual. In this way, it can quickly recognise whether there has been a hacker attack or a technical error.
An illustrative example: In a factory, a system analyses a hundred machines and knows their usual power consumption patterns. If a machine suddenly starts up at night when it should actually be off, this is recognised immediately. In this way, damage or attacks can be prevented and processes optimised.
Pattern-based anomaly detection therefore helps companies to minimise risks and make their processes safer and more efficient.
Deutsch 
English (UK)