摘要
比较了多种类型的核函数下倒向随机微分方程(BSDE)中生成元z的非参数估计方法,利用不同的核函数估计BSDE中的生成元z的非参数估计,在均方误差意义下比较了8种不同的核函数下得到的BSDE的生成元z的非参数估计的精度,统计分析结果显示Gaussian核函数下的估计效果最好。
The comparison of several kernel functions in nonparametric estimation of generator z of backward stochastic differential equation(BSDE) are discussed in this paper.Utilizing different kernel functions for nonparametric estimation of generator z of BSDE,the estimation accuracies of 8 kernel functions in terms of mean squares error(MSE) are investigated,the experimental results with statistical methods show that the estimation effectiveness of Gaussian kernel function is the best.
出处
《统计与信息论坛》
CSSCI
2011年第11期8-12,共5页
Journal of Statistics and Information
关键词
倒向随机微分方程
随机微分方程
非参数估计
核函数
均方误差
backward stochastic differential equation
stochastic differential equation
nonparametric estimation
kernel function
mean square error