An Approach to Isolating Sources of Errors in Invalid State Space Models based on Stochastic Approximation

This study presents a methodology for isolating sources of misspecification in state space models that are known to be invalid. The methodology relies on a technique based on stochastic approximation in the context of a Bayesian formulation. This approach has significant advantages in computational efficiency, relative to a straightforward Bayesian analysis, in large-scale systems. Moreover, it applies to arbitrary misspecified parameters in the model under consideration.