The solution of the Kolmogorov backward equation is expressed as a functional integral by means of the Feynman–Kac formula. The expectation value is approximated as a mean over trajectories. In order to reduce the variance of the estimate, importance sampling is utilized. From the optimal...

This paper deals with the identification and maximum likelihood estimation of the parameters of a stochastic differential equation from discrete time sampling. Score function and maximum likelihood equations are derived explicitly. The stochastic differential equation system is extended to allow...

The unknown structural parameters of a continuous/discrete state space model are estimated by maximum likelihood in the presence of irregular sampling, missing values, and cross-sections of time series (panel data). Exogenous (control) variables are included, and the sampling scheme and missing...