312512 Hidden Markov Models (V) (SoSe 2025)

Hidden Markov models (HMMs) are time series models that are useful for analysing sequential data that are driven by underlying states. Typical examples are animal movement data (where the underlying states represent the animal's motivations), medical records (where the states could correspond to different disease stages), and financial data (where the states are associated with different levels of nervousness of the market).

The course will formally introduce the class of HMMs and the associated inferential methods (parameter estimation, state decoding, prediction, model selection, model checking). Participants will be taught how to code HMMs in R from scratch. We will also be discussing various closely related classes of models. And of course there will be lots of examples throughout.