scientific article; zbMATH DE number 3928097

From MaRDI portal
Publication:3702284





Non-parametric kernel type estimator of a density function of a random variable \(\xi\) is the statistic \[ a_ nn^{-1}\sum^{n}_{i=1}K(a_ n(x-X_ i)),\quad x\in {\mathbb{R}}, \] where \(X_ 1,X_ 2,...,X_ n\) is a sample of \(\xi\), K is the kernel function and \(\{a_ n,n\geq 1\}\) is a sequence of real numbers. The pioneering works in the study of kernel estimates of M. Rosenblatt and E. Parzen were concerned almost exclusively with bias and variance asymptotics. In this book the author presents his subsequent works on kernel estimates of density functions and their derivatives and of regression curves. The book consists of five chapters. Chapter 1 is concerned with asymptotics of the mean integrated square error of kernel density estimates and of the mean square error of kernel estimates of density derivatives in a point. Chapter 2 contains theorems on strong consistency of kernel estimates in the spaces C(\({\mathbb{R}})\), \(L_ 2({\mathbb{R}})\), and some related results. Chapter 3 contains results on deviation estimations of asymptotical distributions. In chapter 4 the author studies the asymptotical properties of kernel estimates of regression curves. Chapter 5 is devoted to N. N. \(\hat Ce\)ncov type estimates of density functions.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3702284)