PERBANDINGAN REGRESI METODE ROBUST DENGAN METODE OLS STUDY KASUS PENGARUH INFLASI DAN PDRB TERHADAP PENGANGGURAN TERBUKA DI PROVINSI JAWA TENGAH

Abstract

The least squares method (Ordinary Least Square = OLS) is a widely used estimation method for estimating regression model parameters. This method has assumptions that some of them in real data use often can not be met. If there is sine then the least squares method is inaccurate to estimate the parameters. To solve this problem, one of the methods used is robust regression method. Robust regression was introduced by Andrews (1972) and is a regression method used when the distribution of abnormal error and or some outliers influences the model (Ryan, 1997). The data in this study will compare which model is the best OLS or Robust on Penganngura in Central Java province 2009.
Variables used are unemployment as dependent variable and GRDP, Inflation as
Independent variable. The result shows that all significant variables to the unemployment variable will be teteapi data from both models both OLS and show is not normal. But when compared with the OLS model, Robust method is better that R-squared shows as 16.65% and OLS shows 15.56%.

Keywords: OLS, Robust Regression, Inflation, GRDP, Unemployment