PERBANDINGAN HASIL MODEL REGRESI ROBUST ESTIMASI M (METHOD OF MOMENT), ESTIMASI M (MAXIMUM LIKELIHOOD TYPE), DAN ESTIMASI LTS (LEAST TRIMMED SQUARE) PADA PRODUKSI PADI DI KECAMATAN SEKARAN

Abstract

Regression analysis is a statistical method used to determine the effect of the dependent variable on the independent variable. The aim of regression analysis is to obtain an estimated model of regression model parameters from data. One of the methods used to estimate regression parameters is the MKT method (Least Squares Method). This method is not appropriate to use on data that contains outliers. Therefore, we need an alternative method that is robust to the presence of outliers, namely robust regression. In this study, the robust method used is robust regression, MM estimation, M estimation, and LTS estimation. The aim of this research is to compare the three estimation methods and select the best estimation model based on the coefficient of determination and mean square error. The case study in this research is rice production data in Sekaran sub-district with the dependent variable being rice production, the independent variables land area, productivity and population. The results of the research show that the Least Trimmed Square (LTS) robust regression method is the method that produces the best model, because the Least Trimmed Square (LTS) method has a greater determination value and a smaller Mean Square Error (MSE) compared to the MM estimation and M estimation methods.