This question already has an answer here:

I would like to detect anomalies in a time series data using a moving average approach. Basically, the system will rise an alarm when a data point is far away (3*std for example) from the current mean.

The problem is that the data contains periodic deviations and I do not want to rise alarms for these periodic deviations (even if they are far away from the mean). In other words I would like to rise alarms only at non-periodic deviations. What basic approaches I can try to achieve this?

Thank you. Mario


marked as duplicate by Rob Hyndman, Andy W, gung - Reinstate Monica, John, Silverfish Aug 7 '15 at 7:10

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

Browse other questions tagged or ask your own question.