Bootstrap Kernel Estimators For Multivariate Density Functional

摘要

Let X1,…,Xn be independent identically distributed random variable series, having the common probability density function. The bootstrap method is discussed for the kernel estimation f(AT) of the density function fn(X)=1hd det(S)1/2∑K(X-X1)TS-1(XX1)/h2). It has been shown that the bootstrap works for kernel estimates.