Efficient Bayesian estimation for GARCH-type models via sequential Monte Carlo

This thesis develops a new and principled approach for estimation, prediction and model selection for a class of challenging models in econometrics, which are used to predict the dynamics of the volatility of financial asset returns. The results of both the simulation and empirical study in this research showcased the advantages of the proposed approach, offering improved robustness and more appropriate uncertainty quantification. The new methods will enable practitioners to gain more information and evaluate different models' predictive performance in a more efficient and principled manner, for long financial time series data