1: \begin{abstract}

2: This paper proposes a widely applicable method of approximate

3: maxi\-mum-likelihood estimation for multivariate diffusion process from

4: discretely sampled data. A closed-form asymptotic expansion for

5: transition density is proposed and accompanied by an algorithm

6: containing only basic and explicit calculations for delivering any

7: arbitrary order of the expansion. The likelihood function is thus

8: approximated explicitly and employed in statistical estimation. The

9: performance of our method is demonstrated by Monte Carlo simulations

10: from implementing several examples, which represent a wide range of

11: commonly used diffusion models. The convergence related to the

12: expansion and the estimation method are theoretically justified using

13: the theory of Watanabe [\textit{Ann. Probab.} \textbf{15} (1987) 1--39]

14: and Yoshida [\textit{J. Japan Statist. Soc.} \textbf{22} (1992)

15: 139--159] on analysis of the generalized random variables under some

16: standard sufficient conditions.

17: \end{abstract}