FUNGSI PEMBOBOT TUKEY BISQUARE DAN WELSCH PADA REGRESI ROBUST ESTIMASI-S

Abstract

The research studied the robust regression estimation-S on Tukey-Bisquare and Welsch weighting function. Due to the data contains outliers, we therefore used robust regression to handle outliers. The purpose is to find the best model between the Tukey Bisquare or Welsch. The mean square error (MSE) and adjusted is then used. The simulation data is a human development index (HDI) in Papua in 2021. The result showed that the robust regression model on the S-estimation between Tukey Bisquare and Welsch are similar (close). However, the adjusted weighting function Welsch is greater than the Tukey Bisquare, and the MSE of the Welsch is smaller than the Tukey Bisquare, so we recommended the robust S-estimation Welsch is a little better.