ABSTRACT
This chapter describes methods for estimating parameters in stochastic differential equations. It introduces a very general maximum likelihood method for parameter estimation in continuous/discrete-time state space models. There are at least three types of competing algorithms that are computationally efficient. The most general, and also the computationally least efficient class of algorithms, are the Monte Carlo simulation based estimators. The Fokker-Planck based estimators are computationally more efficient (at least in low dimensions). Finally, the series expansion approach is the computationally most efficient algorithm but it is also the most restrictive estimator. The chapter discusses the generalized method-of-moments (GMM) and introduces the moment restrictions in formulating the GMM method.
