Mitigating consequences of the Markov property
Talk, Colloquium of the center of Statistics, Bielefeld University, Bielefeld, Germany
I presented the results of my master thesis which focussed on methods for mitigation the undesired consequences of the Markov assumption in hidden Markov models. In particular, I derived properties of the state process of periodically inhomogeneous Markov chains, and extended the model class of hidden semi-Markov models to allow for inhomogeneity in the state dwell-time distributinos. You can download the slides here