On Hybridizations of Fourth Order Kernel of the Beta Polynomial Family.

Israel Uzuazor SILOKO, Osayomore IKPOTOKIN, Edith Akpevwe SILOKO

Abstract


The usual second order nonparametric kernel estimators are of wide uses in data analysis and visualization but constrained with slow convergence rate. Higher order kernels provide a faster convergence rates and are known to be bias reducing kernels. In this paper, we propose a hybrid of the fourth order kernel which is a merger of two successive fourth order kernels and the statistical properties of these hybrid kernels were study. The results of our simulation reveals that the proposed higher order hybrid kernels outperformed their corresponding parent’s kernel functions using the asymptotic mean integrated squared error.


Keywords


Kernel, Bandwidth, AMISE, Beta Polynomial, Higher Order Kernel.

Full Text:

PDF


DOI: http://dx.doi.org/10.18187/pjsor.v15i3.2625

Refbacks

  • There are currently no refbacks.




Copyright (c) 2019 Pakistan Journal of Statistics and Operation Research

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Title

On Hybridizations of Fourth Order Kernel of the Beta Polynomial Family.

Keywords

Kernel, Bandwidth, AMISE, Beta Polynomial, Higher Order Kernel.

Description

The usual second order nonparametric kernel estimators are of wide uses in data analysis and visualization but constrained with slow convergence rate. Higher order kernels provide a faster convergence rates and are known to be bias reducing kernels. In this paper, we propose a hybrid of the fourth order kernel which is a merger of two successive fourth order kernels and the statistical properties of these hybrid kernels were study. The results of our simulation reveals that the proposed higher order hybrid kernels outperformed their corresponding parent’s kernel functions using the asymptotic mean integrated squared error.


Date

2019-09-13

Identifier


Source

Pakistan Journal of Statistics and Operation Research; Vol. 15 No. 3, 2019



Print ISSN: 1816-2711 | Electronic ISSN: 2220-5810