Original scientific paper
https://doi.org/10.17535/crorr.2018.0025
An Improved Robust Regression Model for Response Surface Methodology
Efosa Edionwe
; Department of Mathematical Sciences, Edwin Clark University, Delta State, Nigeria
J. I. Mbegbu
; Department of Statistics, University of Benin, Benin, Nigeria
N. Ekhosuehi
; Department of Statistics, University of Benin, Benin, Nigeria
H. O. Obiora-Ilouno
; Department of Statistics, Nnamdi Azikiwe University, Awka, Nigeria
Abstract
In production, manufacturing and several other allied industries, appropriate tool is applied in the analysis of data in order to enhance the opportunity for product and process optimization. A statistical tool that has successfully been used to achieve this goal is Response Surface Methodology (RSM). A recent trend in the modeling phase of RSM involves the use of semi-parametric regression models which are hybrids of the Ordinary Least Squares (OLS) and the Local Linear Regression (LLR) models. In this paper, we propose a modification in the current structure of the semi-parametric Model Robust Regression 2 (MRR2) with a view to improving its sensitivity to local trends and patterns in data. The proposed model is applied to two multiple response optimization problems from the literature. The results of goodness-of-fits and optimal solutions confirm that the proposed model performs better than the MRR2.
Keywords
desirability function; genetic algorithm; local linear regression; multiple response optimization problem; semi-parametric regression models
Hrčak ID:
212397
URI
Publication date:
13.12.2018.
Visits: 1.347 *