Nonclassical parameters in kernel estimation
Journal Title: Bulletin de la Société des sciences et des lettres de Łódź, Série: Recherches sur les déformations - Year 2016, Vol 0, Issue 1
Abstract
In kernel method, using in estimation as well as in hypothesis testing problems, two parameters should be fixed: kernel function and smoothing parameter. Some methods of kernel estimation and methods of choosing kernel parameters (classical and nonclassical) are presented. Basing on simulation study results, chosen kernel estimation methods are compared, taking into account their properties.
Authors and Affiliations
Aleksandra Baszczyńska
Keywords
Related Articles
- EP ID EP191296
- DOI -
- Views 66
- Downloads 0
On special saturated sets / O specjalnych zbiorach nasyconych
In the paper the examples of sets which are ’nice’ in algebraical sense and saturated non-measurable with respect to the Lebesgue measure are considered. It is investigated under what conditions on measurable space those...
Some remarks on similar topologies / Pewne uwagi o podobnych topologiach
Let T1, T2 be topologies defined on a set X. Assume that T1 and T2 are similar, that means the families of sets with nonempty interior in topological spaces (X; T1) and (X; T2) are equal. The aim of our paper is to check...
Heavy ions sputtering and implantation of surface monitored with PIXE / Rozpylanie ciężkimi jonami i implantacja powierzchni monitorowana przy pomocy PIXE
Characteristic X-rays emitted from surfaces and specially prepared thin films as a result of impact of slow heavy ions were measured and analysed. Kinematics of the interaction was simulated numerically with stopping and...
Complete sufficient statistics for Markov chains
This paper presents a complete suffcient statistic for the class of Markov chains with a finite state S for some natural subspaces PZ of the set of all transition probabilities P. This statistic is the random transition...
The bounds of some determinants for functions of bounded turning of order alpha
-