Likelihood based inference for diffusion driven models.

摘要: This paper provides methods for carrying out likelihood based inference diffusion driven models, example discretely observed multivariate diffusions, continuous time stochastic volatility models and counting process models. The diffusions can potentially be non-stationary. Although our are sampling based, making use of Markov chain Monte Carlo to sample the posterior distribution relevant unknowns, general strategies details different from previous work along these lines. we develop simple implement simulation efficient. Importantly, unlike methods, performance technique is not worsened, in fact it improves, as degree latent augmentation increased reduce bias Euler approximation. In addition, method subject a degeneracy that afflicts techniques when increased. We also discuss issues model choice, checking filtering. ideas applied both simulated real data.