Editing Hidden Markov model (section)

=== General state spaces ===
In the hidden Markov models considered above, the state space of the hidden variables is discrete, while the observations themselves can either be discrete (typically generated from a [[categorical distribution]]) or continuous (typically from a [[Gaussian distribution]]). Hidden Markov models can also be generalized to allow continuous state spaces. Examples of such models are those where the Markov process over hidden variables is a [[linear dynamical system]], with a linear relationship among related variables and where all hidden and observed variables follow a [[Gaussian distribution]]. In simple cases, such as the linear dynamical system just mentioned, exact inference is tractable (in this case, using the [[Kalman filter]]); however, in general, exact inference in HMMs with continuous latent variables is infeasible, and approximate methods must be used, such as the [[extended Kalman filter]] or the [[particle filter]].

Nowadays, inference in hidden Markov models is performed in [[Nonparametric statistics|nonparametric]] settings, where the dependency structure enables  [[identifiability]] of the model<ref>{{Cite journal |last1=Gassiat |first1=E. |last2=Cleynen |first2=A. |last3=Robin |first3=S. |date=2016-01-01 |title=Inference in finite state space non parametric Hidden Markov Models and applications |url=https://doi.org/10.1007/s11222-014-9523-8 |journal=Statistics and Computing |language=en |volume=26 |issue=1 |pages=61–71 |doi=10.1007/s11222-014-9523-8 |issn=1573-1375}}</ref> and the learnability limits are still under exploration.<ref>{{Cite journal |last1=Abraham |first1=Kweku |last2=Gassiat |first2=Elisabeth |last3=Naulet |first3=Zacharie |date=March 2023 |title=Fundamental Limits for Learning Hidden Markov Model Parameters |url=https://ieeexplore.ieee.org/document/9917566 |journal=IEEE Transactions on Information Theory |volume=69 |issue=3 |pages=1777–1794 |doi=10.1109/TIT.2022.3213429 |arxiv=2106.12936 |issn=0018-9448}}</ref>