PENERAPAN METODE REGRESI ROBUST ESTIMASI-M UNTUK KASUS DATA PDRB PERKAPITA PROVINSI JAWA TENGAH TAHUN 2021

Abstract

Economic growth and a more even distribution of income are needed to improve social welfare. One of the most important indicators for measuring the economic growth of a region is the growth value of the gross regional domestic product (GRDP). The size of the gross regional domestic product (GRDP) in Central Java can be predicted using a regression model. One of the estimation methods used in linear regression models is the LS method. The use of LS method becomes less appropriate if there is a violation of classical assumptions caused by outliers in the data. The M-estimation robust regression method can be used to solve outliers problem in the data. In this study, the M-estimation robust regression method was applied to the case of district or city GDP per capita data in Central Java Province in 2021. The weighting functions used were the Tukey bisquare weighting function and the Huber weighting function. The adjusted R-squared value and mean squared error (MSE) are used to determine the criteria for the best model. Based on the research results, it can be concluded that robust M-estimation regression with the Huber weighting function is the optimal parameter estimate for determining the best model. This is because the Huber weighting function has a higher adjusted R-squared value than the bisquare Tukey weighting function, and the Huber weighting function has a lower MSE than the bisquare Tukey weighting function