Mathematical model of the unemployment rate with multiple spline regression

Research related to the open unemployment rate mostly only uses multiple linear regression. However, upon closer examination, the relationships between variables exhibit nonlinear patterns. Therefore, in this paper, a model is developed to address these nonlinear patterns. The purpose of this study was to analyze the characteristics and determine the best model for the open unemployment rate (UR) in South Sulawesi Province using truncated Spline nonparametric regression. In this study, the variables used were the labor force participation rate, the percentage of poor people and the average length of schooling using the Spline truncated method which is a model development from nonparametric regression that can estimate data wherever the data pattern moves. The results showed that there were 15 districts/cities that had UR below the average, while 9 other districts/cities had UR above the provincial average. The district/city with the lowest open unemployment rate is Enrekang district at 2,34 percent. Meanwhile, the district/city with the highest open unemployment rate is Makassar City at 13,18 percent. The best nonparametric regression model is to use three knot points with a minimum GCV value of 1,1352. Based on the results of this analysis, it can be concluded that the variable labor force participation rate, the percentage of the poor and the average length of schooling have a significant effect on the open unemployment rate in South Sulawesi Province in 2021 with a coefficient of determination indicating a good model of 95,69%.