No title
To implement maximum likelihood estimation in state-space models, the log-likelihood function must be approximated. We study such approximations based on particle filters, and in particular conditions for consistency of the corresponding approximate maximum likelihood estimator. Numerical results illustrate the theory.
