|
Abstract:
|
In this talk, we propose new smoothing methods for state-space models with unknown parameters. The first approach originates from the particle smoothing algorithm, but with an adjustment in the backward resampling weights. The second one is a new method combining parameter estimation and smoothing for general state-space models. The method is straightforward but highly efficient. Using simulated and real data, we show that this is the best existing method using Sequential Monte Carlo to solve the joint Bayesian smoothing problem.
|
