I am looking for a comprehensive list of regime switching/change point models/techniques which can be used to model different regimes / change points in financial time series. What I found so far are:

Regime Switching:

  • Hidden Markovian Regime Switching (HMRS)
  • Interactive Hidden Markovian Regime Switching (IHMRS)
  • Self Exciting Threshold Autoregressive (SETAR)
  • ....

Change Point Analysis:

  • Bayesian Change Point Detection
  • Extreme Change Point Detection
  • ...

What other techniques are out there?

PS: I will try to keep the question updated so that it turns into a wiki entry.

  • $\begingroup$ Hi: Kim and Nelson have a text on regime switching. It's kind of old now but I remember it being pretty good at the time ( maybe late 90's ). IIRC, the algorithms are MCMC type so gibbs sampling-metropolis hastings. $\endgroup$ – mark leeds Nov 12 '20 at 2:21
  • $\begingroup$ the same Gibbs who worked with Maxwell and Boltzmann? $\endgroup$ – develarist Nov 12 '20 at 2:45
  • $\begingroup$ @develarist yes, but the name is a misnomer - it's only named after him because the first applications of gibbs sampling were in statistical mechanics. $\endgroup$ – rubikscube09 Nov 12 '20 at 3:52
  • $\begingroup$ Yes, the famous paper whose title escapes me is by gibbs and rosenblath. They were doing something with image restoration in the 50's and using gibba sampling in that framework. The statistics community jumped on it like 30-35 years later in late 80's-early 90's.. $\endgroup$ – mark leeds Nov 12 '20 at 5:01

An Evaluation of Change Point Detection Algorithms might be of use.


Your Answer

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

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