跳-扩散模型中时变参数的核函数加权估计

Kernel Weighted Estimators of Time-varying Parameters for Discretely Observed Jump-Diffusion Model

本文主要研究跳一扩散模型中时变参数的核函数加权估计。基于带复合Poisson跳的扩散模型的离散观测样本,首先得到了漂移参数的核函数加权最小二乘估计及其标准误差,然后利用分位回归方法得到了扩散参数的核函数加权分位回归估计,并证明了所求估计的相合性。最后通过模拟说明了估计量的有效性。

This paper studies the kernel weighted estimators of the time-varying parameters of the jump-diffusion model.First and foremost,based on discretely observed sample of stochastic differential equation model with compound Poisson jumps,the kernel weighted least-square estimators of the drift parameters is gained and their standard errors are discussed in our paper.Besides,we find the kernel weighted quantile regression estimator of the diffusion parameter by using the quantile regression method.The consistency of the kernel weighted quantile regression estimator of the diffusion parameter is verified.Finally,it is shown that the estimations are effective through a simulation study.