1

If I have an algorithm for detecting a set of data points that indicate with a high level of certainty that some event has occured OR that behaviors outside of a set model are occuring.

What would you call the algorithm used to detect these anomalies?

And

Would calling it an "analytic" be incorrect?

Jay Hawk
  • 11
  • 1
  • I think it depends on how the method works. If you can not get more specific, I think Anomaly Detection is a proper term. I would be careful with terms like "analytical", since it can give a missleading impression that a method is not approximate. – André Sep 17 '18 at 12:52
  • @André I was trying to avoid being specific on purpose, because the intent is to collect many different methods for detecting many different attacks in a programmatic manner.

    One specific example would be detecting a ICMP tunneling by observing unusual spikes in ICMP messages or unusually large ICMP packets.

     – Jay Hawk Sep 24 '18 at 16:46
  • if event is not acceptable event, u can call ur algo as abnormality detection also. There are many abnormality detection algorithms based on problem statement, It can be supervised/ unsupervised or combination of both. -https://machinelearningstories.blogspot.com/2018/07/anomaly-detection-anomaly-detection-by.html – Arpit Sisodia Oct 01 '18 at 04:10
  • Analytic would be too generalized term for ur algo – Arpit Sisodia Oct 01 '18 at 04:11

0 Answers0