ANALISIS KETEGARAN REGRESI ROBUST TERHADAP LETAK PENCILAN: STUDI PERBANDINGAN

Abstract. The effect of outlier numbers and location on regression coefficient is studied by comparing the estimated regression coefficient and Mean Square Error (MSE) resulted by M-estimator, MM-estimator, and by Ordinary least square (OLS). In order to show the robustness of robust regression methods we used four sample sizes: n = 20, 60,100 and 200 which contaminated by 5%, 10%, 15%, 20%, 25% and 30% outliers located in the center and end of the data. The result shows that MM-Estimator is the most robust method compare to M-estimator and OLS