Overview

Original Research

ROBUST ESTIMATOR FOR FINITE POPULATION TOTAL

AJWANG’ STELLAMARIS ADHIAMBO 1, ROMANUS ODHAMBO OTIENO 2, THOMAS MAGETO 3, and DAVID ALILA 4.

Vol 18, No 02 ( 2023 )   |  DOI: 10.17605/OSF.IO/CUYX7   |   Author Affiliation: Department of Mathematics(Statistics option) Programme, Pan African University, Institute for Basic Sciences, Technology and Innovation(PAUSTI), Nairobi, Kenya 1; Department of Statistics and Actuarial Sciences, Jomo Kenyatta University of Agriculture and Technology, Nairobi, Kenya 2,3; Department of Mathematics, Masinde Muliro University of Science and Technology, Kenya 4.   |   Licensing: CC 4.0   |   Pg no: 305-314   |   Published on: 13-02-2023

Abstract

The kernel regression estimator is a flexible and widely used nonparametric estimator that estimates a regression function. Statistical learning appears to be a promising field in which some algorithms resulting from machine learning are interpreted as statistical methods. Boosting is among the most studied machine learning techniques in this paper, a new improvement of the kernel density regression estimator is proposed, with a target of producing smaller estimates of the finite population total. The study was aimed at estimating the finite population total by incorporating the adaptive boosting technique to the nonparametric regression estimator. A numerical study using a simulated population was conducted in order to evaluate the performance of the proposed estimator and compare it with the existing one. The outcome of the proposed estimator is evaluated and presented. The simulation experiment were very promising; it shows that our modified kernel density estimator performs well in all cases, then the normal kernel density estimator.

Keywords

Nonparametric Estimation, Kernel Density Estimator, Bias Reduction, Adaboost, Finite, Population Total