Bayesian time-series econometrics 2400-ZEWW828
The detailed program. The course will consist of 2 modules:
1) Introduction to the Bayesian inference:
a) Bayes' formula, differences between the Bayesian and frequentist paradigms
b) Linear regression in the Bayesian framework
c) AutoRegression (AR) in the Bayesian framework
d) Markov Chain Monte Carlo (MCMC) methods (Gibbs sampling, several variants of the Metropolis-Hastings algorithm and convergence monitoring)
2) Models discussed (the list of models actually considered during classes may change):
a) Vector AutoRegression (VAR)
b) Structural VAR
c) State-space models
d) Time-Varying Parameters models (TVP)
e) Model described in the article J. Qiu, S. R. Jammalamadaka, and N. Ning (2018), "Multivariate Bayesian structural time series model", Journal of Machine Learning Research, and the dedicated R package "MBSTS"
f) NOTE: there is an option to propose the estimation of any model that participants find useful or interesting, e.g. Stochastic Volatility, Dynamic Factors Models, TVP-VAR, Local Projection, etc.
Estimated student’s workload: 3ECTS x 25h = 75h
(K) - godziny kontaktowe (S) - godziny pracy samodzielnej
wykład (zajęcia): 0h (K) 0h (S)
ćwiczenia (zajęcia): 30h (K) 0h (S)
egzamin: 0h (K) 0h (S)
konsultacje: 0h (K) 0h (S)
przygotowanie do ćwiczeń: 0h (K) 10h (S)
przygotowanie do wykładów: 0h (K) 0h (S)
przygotowanie do kolokwium: 0h (K) 0h (S)
przygotowanie do egzaminu: 0h (K) 0h (S)
przygotowanie projektu: 0h (K) 35h (S)
Razem: 30h (K) + 45h (S) = 75h
Course coordinators
Type of course
Learning outcomes
Knowledge: The student will understand the Bayesian language; will be able to write code in her/his preferred environment; will learn models used in the modern literature.
Skills: How to use the potential of Bayesian methods in economic/data science modeling; An interested and ambitious student has a chance to become the Bayesian thinker and not (only) a Bayesian user.
Assessment criteria
Preparing empirical report in (max.) 2-person teams, which applies Bayesian methods and uses any programming environment, e.g. R, Python, Matlab, Stata, etc.
To pass the classes the student should submit a report and don’t have more than 3 unexcused absences.
Bibliography
Mainly my detailed slides and/or:
Geweke, J. (2006), Contemporary Bayesian Econometrics and Statistics, Wiley.
Hamilton, J. D. (1994), Time Series Analysis, Princeton University Press.
Koop, G. (2003), Bayesian Econometrics, Wiley
Lancaster, T. (2004), An Introduction to Modern Bayesian Econometrics, Wiley-Blackwell.
Zellner, A. (1971), An Introduction to Bayesian Inference in Econometrics, Wiley.