CEU Electronic Theses and Dissertations, 2009
| Author | Lukács, Miklós |
|---|---|
| Title | Bayesian model for forecasting Hungarian inflation: A comparison of the performance of BVAR, VAR and ARIMA models |
| Summary | This thesis aims to conduct forecasts the Hungarian inflation for two years ahead based on observations from January 1998 to March 2007. The forecasting precision of three different models are introduced: ARIMA, VAR and Bayesian VAR. The prior assumptions of the Bayesian model are specified with the Minnesota prior. According to the calculated loss functions this model performs much better than the other two. The VAR model proves to be very poor, because of the large number of estimated parameters. |
| Supervisor | Korosi, Gabor |
| Department | Economics MA |
| Full text | https://www.etd.ceu.edu/2009/lukacs_miklos.pdf |
Visit the CEU Library.
© 2007-2021, Central European University