Adam, Timo: On some flexible extensions of hidden Markov models. 2020
Inhalt
- Table of contents
- Introduction
- Introduction to hidden Markov models
- A brief history of hidden Markov models
- Outline of the thesis
- Statement of contribution and related work
- Gradient boosting in Markov-switching generalized additive models for location, scale, and shape
- Introduction
- Model formulation and dependence structure
- Model fitting
- The MS-gamboostLSS algorithm
- Specification of base-learners
- Choice of the number of boosting iterations
- Selecting the number of states
- Simulation experiments
- Application to energy prices in Spain
- Discussion
- Non-parametric inference in hidden Markov models for discrete-valued time series
- Introduction
- Model formulation and model fitting
- Model formulation and dependence structure
- Likelihood evaluation
- Roughness penalization
- Model fitting and parameter constraints
- Choice of the tuning parameters
- Simulation experiments
- Application to earthquake counts
- Discussion
- Joint modeling of multi-scale time series using hierarchical hidden Markov models
- Introduction
- Model formulation and dependence structure
- Multivariate hidden Markov models
- Hierarchical hidden Markov models
- Incorporating covariates into the model
- Some remarks on model fitting and related topics
- Real-data applications
- Discussion
- Conclusions
- A forward algorithm for likelihood evaluation in hierarchical hidden Markov models
- Estimated coefficients for the fine-scale state transition probabilities
- Bibliography