Parameter estimation by deterministic approximation of a solution of a stochastic differential equation
Parameter estimation by deterministic approximation of a solution of a stochastic differential equation
复制标题
通过随机微分方程解的确定性近似进行参数估计
DOI：
10.1080/15326349908807135
复制
发表时间：
1990
期刊：
影响因子：
--
通讯作者：
R. A. Kasonga
中科院分区：
文献类型：
--
作者：
R. A. Kasonga
考虑Rd中的扩散X=(Xt t≥0)，它是[M0001]和确定性类似物[m0002]的解，我们将证明当f满足通常的Lipschitz条件和增长条件时，(I)和(II)的解在某种意义上是接近的。然后利用这一结果证明了参数θ的最小二乘型估计的相合性。对于模型(I)的一种特殊情况，将证明一个强相合性结果。最后，我们将使用模拟观测来比较我们的估计器与Dorogocev[1]和Prakasa Rao[3]提出的估计器。
Consider a diffusion X = (Xt t ≥ 0) in Rd which is a solution of [m0001] and the deterministic analogue [m0002] We shall prove that when f satisfies the usual Lipschitz and growth conditions, the solutions of (i) and (ii) are close in some sense. Then we use this result to prove consistency of a least squares type estimator of the parameter θ. A strong consistency result will be proved for a special case of model (i). Finally, simulated observations will be used to compare our estimator with the one proposed by Dorogovcev [1] and Prakasa Rao [3].