An extended question to this old post: Does applying ARMA-GARCH require stationarity?

I suspect my data (long data with a few hundred thousand points) is not completely stationary. But some segments look more stable and may be interpreted by stationary processes while the others have "volatility clusters". So how to handle such data, like what tests to use to determine models to be use?

  • $\begingroup$ Volatility clustering does not imply non-stationarity. $\endgroup$ – Matthew Gunn Nov 2 '18 at 21:20
  • $\begingroup$ ok, that resonates what the previous post said. but any recommendation on how to process the data? $\endgroup$ – user3078106 Nov 2 '18 at 23:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.