摘要
运用泰勒展式讨论了非参数回归模型中未知误差分布函数f(e)的核密度估计^fn(e)的渐进性质,以及估计量^fn(e)中光滑参数的选择,并给出了f(e)的置信区间.
Using Taylor expansion, the authors studied some extension asymptotic properties of a nonparametric kernel density estimation fn (e) of an unknown error density function f( e ) in a nonparametrie regression model. Then they studied the choice of smoothing parameters both in the regression function and error density. Finally, an approximation confidence interval of f(e) was given.
出处
《四川大学学报(自然科学版)》
CAS
CSCD
北大核心
2005年第5期905-909,共5页
Journal of Sichuan University(Natural Science Edition)
基金
国家社科基金(01BTJ003)
关键词
非参数回归
误差分布
核密度估计
光滑参数
nonparametric regression
error density
asymptotic unbiased
smoothing parameter