On White Noises Driven by Hidden Markov Chains
Christian Francq, Michel Roussignol · Journal of Time Series Analysis · 1997
We consider a time series model where the variance of the underlying process depends on the state of a non‐observed Markov chain. Maximum likelihood estimates are shown to be consistent. Estimators with asymptotic Gaussian distribution are proposed. Prediction and identification are also mentioned. This is illustrated by means of real and simulated data sets