A0480
Title: A sequential Monte Carlo approach to estimate noncausal processes
Authors: Francesco Giancaterini - University of Rome Tor Vergata (Italy) [presenting]
Gianluca Cubadda - University of Rome Tor Vergata (Italy)
Stefano Grassi - University of Rome 'Tor Vergata' (Italy)
Abstract: A Bayesian estimation approach is introduced for mixed causal and noncausal models using sequential Monte Carlo (SMC) methods in univariate and multivariate contexts. The SMC method allows the estimation of the investigated process by considering different assumptions about the distributions of the error term. Consequently, the SMC approach facilitates the comparison of marginal data densities under different assumptions, helping to identify the error-term assumption that best fits the data. Furthermore, SMC offers extensive parallelization possibilities, significantly reducing estimation time and mitigating the risk of becoming trapped in local minima. Simulation studies demonstrate the strong ability of the algorithm to correctly identify the process and provide accurate estimates.
