Bandwidth Choice for Kernel Density Estimates.
|Field:||Applied statistics, operation research|
|Type:||Article in Proceedings|
|Keywords:||multivariate kernel; bandwidth matrix; AMISE|
Kernel density estimates attempt to reconstruct the probability density from the sample has come using the sample values and a few assumptions as possible. We focus on multivariate density estimates. A main attention is paid to a choice of a smoothing parameter. We extend an idea of an iterative method originally proposed for univariate densities to bivariate ones. The proposed method is compared with cross-validation methods in a simulation study.