Bootstrap methods for generalized autoregressive moving average models

Abstract

This final paper aims to find a suitable Bootstrap Method for the Generalized Autoregressive Moving Average Model. The focus is on the Moving Block Bootstrap (MBB) resampling scheme with its performance being evaluated through a Monte Carlo study and contrasted to their asymptotic Gaussian counterpart. It is stablished that the aforementioned resampling procedure can generate good estimates of parameters bias and confidence intervals. Though, the results rely heavily on the simulated model parameters and block lengths used in the MBB procedure.