I have a 1-min load profile (kW) of a machine (see the figure below). The machine has different operating modes that are evident from the timeseries data. Group1 when the machine is idle, Group2 when running and Group3 when starting.
I have several machines with different profiles (timeseries data) and I need to perform the same analysis. I see that this is unsupervised problem where I need to extract all the available clusters and then label the datapoints depending on the clusters they belong to.
What is the best way to extract the underlying groups/clusters present in my timeseries data?
