Sigap Abror Falah, Trianingsih Eni Lestari, Ramdhan Fazrianto Suwarman
Regression analysis is a statistical method used to understand the relationships between interrelated variables across various sectors. This study compares the effectiveness of two estimators in robust regression models, namely the Maximum-Likelihood (M) and Generalized Scale (GS) estimators, to model the Community Welfare Development Index in the regencies/cities of East Java using data from 2018. The comparison results indicate that the robust regression model with the GS estimator provides better outcomes than the M estimator, with lower AIC and SIC values of 89.260 and 93.464, respectively. The GS estimation results show that the average length of schooling, life expectancy, infant mortality rate, percentage of healthy houses, and percentage of poor population significantly influenced the Community Welfare Development Index of East Java in 2018 by 92.08%, with the remaining 7.92% explained by other factors. Meanwhile, the population factor did not pass the multicollinearity test. However, the comparison results between these two estimators do not lead to the conclusion that the GS estimator is generally better than the M estimator. This is because the effectiveness of each estimator may vary depending on the dataset used. This study differs somewhat from previous research, as it employs Principal Component Analysis to address multicollinearity issues with the aim of minimizing the elimination of independent variables, allowing for a more detailed analysis of each variable. © 2025 American Institute of Physics Inc.. All rights reserved.
Department of Mathematics, Faculty of Mathematics and Natural Sciences, Universitas Negeri Malang, Malang, Indonesia
