312512 Hidden Markov Models (V) (WiSe 2015/2016)

Hidden Markov models (HMMs) constitute a class of stochastic time series models that is relatively easily accessible, yet versatile and rich in mathematical structure. Corresponding models have been successfully applied to a wide range of types of data collected in numerous application fields, including economics, medicine, biology, robotics and sociology.

This lecture course will introduce the basic inferential machinery, including model formulation, parameter estimation, model selection, model checking and state decoding. There will be a strong focus on the practical use of HMMs, with the methods being illustrated using real data examples and associated R code. We will also discuss various ways to extend the basic model structure (covering, e.g., also Markov-switching regression models and state-space models).

Requirements for participation, required level

It will be expected that participants are familiar with basic statistics (e.g. from introductory courses in statistics, stochastics or econometrics). Basic knowledge of the free software R would be advantageous, but is not required.

Bibliography

Zucchini, W. and MacDonald, I.L., Hidden Markov Models for Time Series: An Introduction Using R, Chapman & Hall/CRC press, 2009.