Estimasi Nonlinear Least Trimmed Squares (NLTS) pada Model Regresi Nonlinier yang Dikenai Outlier

Abstract

Constant Elasticity of Substitution (CES) production function is the intrinsic nonlinear regression models that are often used to estimate the data in an industry. Intrinsic nonlinear regression model is a kind of nonlinear regression that can not be linearized, so as to estimate the beta parameters nonlinear statistical model used was Nonlinear Least Squares (NLS) using a first order taylor series approach used in the Gauss Newton iteration. One of the problems often encountered in the analysis of data is an outlier, the presence of outliers in the data analysis greatly influence the results of the analysis so it becomes less valid and the estimation become biased. One method that is resistant to outliers regression is a method of Nonlinear Least Trimmed Squares. This research aims to determine the characteristics of parameter CES production function which contains outlier. The result shows that parameter of the production function CES which contains outliers are bias, inconsistent. So the CES production function which does not contain outliers better than the are contains outliers.

Keywords

• nonlinear statistical model
• parameter estimation
• constant elasticity of substitution (ces) production function
• outliers
• nonlinear least trimmed square (nlts) method, gauss newton iteration