Journal of Statistics Applications & Probability
Abstract
Multivariate analysis of data is of wide applicability in data science especially in big data analytic due to the volume of concealed information to be analyzed. Accurate analysis of multivariate variables is pertinent because predictions from analyzed data are good statistical indicators for making helpful decisions economically and industrially. One of the statistical analytic tools for analyzing multidimensional observations is the kernel density estimator in data exploration and visualization. The functionality of the kernel depends on the kernel function and bandwidth which influences smoothness of estimates. Several kernel functions and bandwidth selectors exist in literature; however novel estimators are being introduced to handle complex circumstances. This paper introduces a new multivariate beta kernel functions whose derivation is contingent on the product techniques. The performances of the newly introduced and existing kernels are evaluated with a known objective function and the numerical results distinctly indicating that the introduced family transcended the traditional beta family.
Digital Object Identifier (DOI)
http://dx.doi.org/10.18576/jsap/120340
Recommended Citation
U. Siloko, I.; A. Siloko, E.; A. Ojobor, S.; S. Awad, Wasan; Enoyoze, E.; and C. Ishiekwene, C.
(2023)
"A New Multivariate Product Kernel Functions of the Beta Polynomial Family,"
Journal of Statistics Applications & Probability: Vol. 12:
Iss.
3, Article 40.
DOI: http://dx.doi.org/10.18576/jsap/120340
Available at:
https://digitalcommons.aaru.edu.jo/jsap/vol12/iss3/40