The term "anomaly detection" is particularly important in the areas of big data and smart data, artificial intelligence, industry and Factory 4.0. It describes methods used to automatically detect deviations from normal patterns or behaviour in large amounts of data.
Imagine thousands of machine hours are recorded every day in a modern factory. If a machine suddenly runs differently than usual - for example due to an unusually high temperature or an unexpected stop - the system recognises this deviation immediately. In technical jargon, this deviation is called an "anomaly". Anomaly detection helps to detect errors, attempted fraud or cyberattacks at an early stage and thus prevent damage or failures.
At its core, programmes analyse the collected data for anomalies that were not foreseeable. Anomaly detection also works outside of industry: for example, in online banking when unusually high amounts are transferred and the system triggers an alarm.
For companies, anomaly detection means greater security and the ability to react quickly to problems. It is a modern tool for significantly better controlling risks in digital processes.
Deutsch 
English (UK)