Maximum Likelihood Estimation of Discretely Sampled Diffusions: A Closed‐form Approximation Approach

摘要: When a continuous-time diffusion is observed only at discrete dates, in most cases the transition distribution and hence likelihood function of observations not explicitly computable. Using Hermite polynomials, I construct an explicit sequence closed-form functions show that it converges to true (but unknown) function. document approximation very accurate prove maximizing results estimator maximum shares its asymptotic properties. Monte Carlo evidence reveals this method outperforms other schemes situations relevant for financial models.