APLICATION OF M-ESTIMATION FOR RESPONSE SURFACE MODEL WITH DATA OUTLIERS

Abstract

Relationship between the response variable and the independent variables in a limited area of operation, in the Response Surface Methodology, used a second order polynomial function. The parameters of the model are usually estimated by Least Squares Method. However, this method is very sensitive to outliers. Outliers can affect the results of statistical analysis, as outliers are very likely to produce a large residual and often affect the regression models generated. Thus, the resulting model estimates to be biased and result in errors in the determination of the actual optimal point. Therefore, it takes a strong response surface model/robust against outliers. Proposed as an alternative to using the M-Estimation, for estimating the parameters in the response surface model. In this paper will be shown in the application of M-Esimation Response Surface Model. Keywords: Response Surface Models, Estimation M, Outliers