MCMC System Identification
Background
Maximim likelihood methods are commonly used in parameter estimation of dynamic systems. A limitation of these methods is that the quantification of the estimation error generally relies on expressions that are asymptotic in the data length. As a consequence, there is no good knowledge of the estimation uncertainty for short data lengths.
