I am trying to classify different kinds of time series, but find myself missing good vocabulary. I don't mean that, for a given time series, I try to classify its datapoints into clusters or the like. Instead, I am looking for categories into which I can sort different time series. Two time series belonging to the same category should be apt for similar time series analysis.

For instance, in some processes measured you find that essentially only two values are assumed, the lower is often zero. I like to call this kind of time series "on-off-time series" in my head.

Another category would be continuous, periodic time series like the measured temperature.

Is there an "official" name for these categories? Does the theory of classifying time series have a more distinguishable (than "classifying time series") name which I don't know yet?

  • $\begingroup$ A system of such categories is known as an "ontology". You could try searching for that. $\endgroup$
    – Jon Nordby
    Mar 25, 2023 at 13:29

1 Answer 1


One option is flexible dynamic time warping for time series classification. Dynamic time warping (DTW) is a way of measuring the similarity between two different time series sequences accounting for different speeds. In the example of temperature, DTW could find similar processes with different rates of change.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.