核密度估计的局部积分方差的渐近表示

An asymptotic representation of local ise for kernel density estimation

在非参数统计推断中 ,首选方法是核密度估计 ,但通常很难得到精确表达式 .利用局部交叉生效方法 ,获得随机删失下核密度估计的局部积分方差的渐近表示 ,所得结果改进和推广了Watson ,Leadbetter和Csor go 。

With local cross validation, an asymptotic representation of local integrated squared error for the kernel density estimation under right censorship was obtained. The result in this paper extends and improves the corresponding results presented by Watson, Leadbetter′s and Csorgo, Horvath′s.