Hidden Markov Models for Time Series ? An Introduction Using R

摘要

Hidden Markov models (HMMs) are mixture models in which the mixture component gen-erating an observation is determined by the state of a hidden Markov process instead of astatic mixing distribution. The marginal distribution of an HMM is a mixture distribution.The components of the mixture can be some of the known probability distributions, and theMarkov process can be first or higher order.