[D] Anomaly detection - deviation from normal

Hi everyone,

I am collecting temperature data from a sensor measuring the temperature of a process. The temperature raises to go in steady-state during operation and it cools down when the process is stopped. I would like to set a simple threshold notifying me when the temperature during operation is higher or lower than expected, how should I tackle the following:

* Filter temperatures when the process is warming up (non-steady state)
* How much data should I use to …

anomaly anomaly detection detection deviation machinelearning normal